4

DELMARVA POWER& LIGHT COMPANY

DEVELOPMENT OF A DISPATCHABLE

PV PEAK SHAVING SYSTEM

PV:BONUS PROGRAM - PHASE 1 REPORT

VOLUME 1

PREPARED FOR THE US DEPAR~ OF ENERGYCOOPEILW’IVE AGREEMENT NO. DE-FC36-93CHI0569

October 1995

DISCLAIMER

This report was prepared as an account of work sponsoredby an agency of the United States Government. Neitherthe United States Government nor any agency thereof, norany of their employees, make any warranty, express orimplied, or assumes any legal liability or responsibility forthe accuracy, completeness, or usefulness of anyinformation, apparatus, product, or process disclosed, orrepresents that its use would not infringe privately ownedrights. Reference herein to any specific commercialproduct, process, or service by trade name, trademark,manufacturer, or otherwise does not necessarily constituteor impiy its endorsement, recommendation, or favoring bythe United States Government or any agency thereof. Theviews and opinions of authors expressed herein do notnecessarily state or ref Iect those of the United StatesGovernment or any agency thereof.

Portions

DISCLAIMER

of this document may be illegiblein electronic imageproduced from thedocument.

products. Images arebest available original

De&nmmPowef&ughtcompany PV:BDNUSFhese1Find ReportCoopem#veAgmementNo.DE-FC36-93CNI0669

TABLE OF CONTENTS

VOLUME 1

EXECUTIVE SUMMARY i

INTRODUCTION 1

TASK 4- TEST SYSTEM DEPLOYMENT AND TESTING 1

PROJECTDESCRIPTION 1

PROJECTCOSTSUMMARY 6

PERFORMANCEOFTHEORIGINALSYSTEM 7

TASKS 1 & 2- CONCEPTUAL DESIGN REFINEMENT 14

FUNDAMENTALDESIGNGOALS 14

SYSTEMDESIGNISSUES 15

PERFORMANCEOFTHEMODIFIEDSYSTEM 18

RESOURCEANDSYSTEMSIZINGANALYSIS 22

PEASE2 SYSTEMCONCEPTUALDESIGN 27

TASK 3- MARKET ASSESSMENT 33

APPLIEDENERGYGROUPMARKETRESEARCH 33

COMPETINGTECHNOLOGIESANDVALUE-ADDEDFUNCTIONS 34

TASK 5- PRODUCT DEVELOPMENT AND TESTING PLAN 35

DEPLOYMENTANDTESTINGPLANSFORPEASE2 35

APPENDICES

APPENDIXB - EMERGENCYPOWERASANADDITIONALVALUEFROMDISPATCHABLEPVPEAKSE.AV’JNG

APPENDIXC- ASSESSMENTOFPOLICY,REGULATORY,ANDINSTITUTIONALBARRIERSTOTEE COMMERCIALIZATIONOFA DISPATCHABLEPVPEAKWAVINGSYSTEM

DdnwvaPowu&ugltltcanpaIIy PV:BONUSPhase1Find ReptwtCCWpwativeAWWWM No.DE—PC3M3Wi066V

Executive Summary

INTRODUCTION

This report summarizes the work performed by Dehnarva Power &Light and its subcontractorsin Phase 1 of the US Department of Energy’s PV:BONUS Program. The purpose of the programis to develop products and systems for buildings which utilize photovoltaic (N) technology.Beginning with a cooperative research effort with the University of Delaware’s Center for Energyand Environmental Policy Research Delmarva Power developed and demonstrated the concept ofDispatchable PV Peak Shaving. This concept and the system which resulted horn thedevelopment work are unique from other grid-connected PV systems because it combines a PV,battery energ storage, power conversion and control technologies into an integrated package.

Phase 1 began in July 1993 with the installation of a test and demonstration system at Dehnarva’sNorthern Division General Office building near Newarlq Delaware. Following initial testingthroughout the summer and fdI of 1993, significant modifications were made under anamendment to the DOE contract. Work on Phase 1 concluded in the early spring of 1995.

Significant progress towards the goal of commercializing the system was made during Phase 1,and is summarized below. Based on progress in Phase 1, a proposal to continue the work inPhase 2 was submitted to the US DOE in May 1995. A contract amendment and providing fimdsfor the Phase 2 work is expected in July 1995.

TECHNICAL DEVELOPMENT

The concept of combtig PV and energy storage is not new. In fact, for years small PV systemshave provided power for numerous applications at sites remote from utility service. Storagecapacity for these. off-grid applications is typically large in comparison to the PV array capacity.This permits the user to “ride out” long periods with little or no sun. During the Summer of1992, Delmarva Power and the University of Delaware Center for Energy and EnvironmentalPolicy (CEEP) investigated the use of small PV arrays combined with thermal storage. Results ofthe research indicated that a relatively small amount of storage could have a profound effect onthe value of energy from a PV array. This additional value is derived from the fact that mostsummer-peaking utilities experience their peak loads well after the output of a PV array peaks.This initial research at the University of Delaware, combmed with solar resourcehility loadcoincidence stut~es at SUNY Albany, resulted in three significant conclusions:

1. Utility loads during the summer months are positively correlated with the availabtity ofthe solar resource inmost areas of the United State~

2. A relatively small amount of energy storage permits the output of a PV array to be storeduntil it is most needed, usually during the late afternoon houry and,

Octobtw1996 PageI

Debnnv8Power&ughic8nnpmy~4 rwmentNo.DE-FC36-93CHf0669

PV:BONUSP&se 1Find R@

3. Storage also allows a significant “leveraging effect” in which the output of the system atthe time of peak can be as much as two times the rated capacity of the PV array.

Later analysis indicated that PV systems which included storage (hence Dispatchable PV PeakShaving System) are considerably closer to economic break-even than “non-dispatchable” PVsystems for grid-connected applications, based purely on the value of capacity and energy. Whenemployed as “dual purpose” systems (e.g., to provide both peak shaving and emergency power),the economic value can increase substantially (see Appendix B). The relationship between theutility and customer can also have a significant impact on the economic value of the system.Alternative service arrangements, in which both the utility and customer may realize economicbenefits, can be developed to help commercialize grid-connected PV systems (see Appendix C).

To demonstrate and develop the Dispatchable PV Peak Shaving System Delmarva installed asystem consisting of a 15 kW PV array, 25 kwh of battery energy storage, and a 32 kW inverter.The battery storage and inverter components were provided as an integrated package.Throughout the Summer of 1993, the system was dispatched consistently during peak periods,and verified the association between the availability of solar energy and utility peak loads. Thedemonstration system also provided a test bed for hardware development. System testing andevaluation resulted in considerable enhancements to the system’s controls, operational flexibtity,packaging and safkty. A “user iliendly” interface eliminated the need for a dedicated, externalcontrol computer. PV array conversion efficiency was enhanced by instahg a DC-to-DCconverter at the array/battery module interfhce. Finally, a proposed modular roof-mountingsystem was designed for application with fbture units.

Phase 2 of the PV:BONUS Program will deploy four of the refined Dispatchable PV PeakShaving Systems at host sites across the country for additional testing and evaluation. Thesesystems will provide the pre-commercial operating experience necessary to move forward with anexpanded commercialization effort.

MARKET OVERWEW

Applied Energy Group, Inc. was contracted to analyze the potential market for the DispatchablePV Peak Shaving System among electric utility commercial customers. This was accomplishedprimarily through surveys of customer and trade groups to evaluate potential barriers, and aquantitative analysis of market potential. The details of the market research are explained in aconfidential attachment Appendix A. Not surprisingly, the analyses indicated that the mostsignificant barrier to acceptance is the vexy long payback period for systems at today’s prices. ,However, a small but signitkant market currently exists among “early adopters.” If this marketcould be consolidated in the United States, there is probably sufficient volume to drive theDispatchable PV System’s production costs down substantially. A prelimimuy analysis of theinternational market also revealed applications for the system in areas with high electricity costs.In several cases in the international market, the system already appears to be cost-effective.

Octobsw1996 PageH

Ddm8w8Powu& Lightcomp8ny PV:BONUSPhase1Find Repatt~Ag mmmntNo.DE-FC35-93CMM$9

Overall, the potential domestic market for the Dispatchable PV Peak Shaving System amongcommercial customers is ve~ large. Approximately 14 percent of commercial customers (roughly800,000 buildings) have characteristics indkating a high adoption potential at break-even systemcosts. Obviously, actual market penetration will be less, depending on a number of technical,commercial and marketing issues.

Several other potential markets exist, but were not examined in detail in Phase 1. These includeoff-grid applications, such as village power and remote residential systems, and variousemergency power and back-up systems. These markets require an adaptation of the basicDispatchable PV Peak Shaving System which will be accomplished during Phase 2. Larger”versions of the basic unit can also be used in utility distributed generation applications atsubstations and along heaviIy loaded f~ers.

BUSINESS DEVELOPMENT OVERWEW

The overall goal of the program is the formation of a sustainable business around thedevelopment, integratio~ distribution and servicing of the Dispatchable PV Peak Shaving Systemand its derivatives. Substantial progress has been made towards securing alliances, entering earlymarkets and ident@ing the necessmy resources. These are discussed in a confidential attachment,Appendix D.

October19s6 pageiii

DefnwvePowwa Ughtcompany PV:BOUUSPhaeei Find ReportCoopemtiveAgreementNo.DE-Fc3s43cHIOSS9

Introduction

The conceptual design of the Dkpatchable PV Peak Shaving System was initially developed byDehnarva Power and the Center for Energy and Environmental Policy (CEEP) in late 1992. Theconcept evolved from research done throughout the Summer of 1992 on the use of PV as ademand-side management (DSM) tool. This research indicated that PV, combined with energystorage, was of greater economic value than PV alone for grid-interactive systems. This increasedeconomic value is derived ftom the ability to control the output of a PV system with storage,allowing its use at the time of utility system peaks, local substation or distribution peaks and/orcustomer peaks.

While the design of the system described in this report utilizes Z@tery energy storage, the overallconcept of a dkpatchable PV system is not linked intrinsically to battery technology. In fact, theoriginal research conducted by CEEP and Delmarva Power utilized a combination of PV and hotwater storage to demonstrate the concept.

By the time the PV:BONUS contract was awarded to Delmarva Power in July 1993, the originalconcept had been developed to the point of deploying a test and demonstration system utiliigbattery energy storage. Because the desi~ installation and operation of the test system had amajor influence on subsequent worlq this report begins by describing the work pefiormed underTask 4 of the original PV:BONUS Statement of Work.

Task 4- Test System Deployment and Testing

The major work items under this task were the installation of the test syste~ an accounting of thepedormance and costs of the syst~ a sumnuuy of the technical issues resuking from the testingof the system and recommendations for Phase 2 development and testing.

PROJECT DESCRIPTION

Design and Construction Ovemiew

Delmarva Power constructed a dispatchable photovoltaic (PV) test and demonstration facility atthe Northern DNision General Office Building (NDGO) near Newarlq Delaware during the earlysummer of 1993. The purposes of this facility were to:

1. Gain experience with PV technology

2. Gather solar resource da~,

3. Explore the potential uses of photovoltaics and battery ener~ storage fordispatchable building peak shaving systems; and

October1996 Page1

DehnawePower& ughtcompluly PV:BONUSP/mee1RIM!RqmttCcmpetailveAgreementNo.DE-PC2S42CNIOS69

4. Serve as the test-bed for the development of modular, dispatchable PV systems forbuildings under the Department of Energy’s PVBONUS Program.

The system was later modfied to incorporate some of the improvements identifkd during earlytesting. Detailed engineering for the project began in January 1993 in a collaborative effort whichincluded Delmarva Power, Ascension Technology of Walthaq Massachusetts and AstroPower ofNewark Delaware. The system design phase was followed by the award of separate contracts forsystem integration electrical installatio~ mechanical installatio~ and data acquisition systemequipment. Site construction began with preliminary electrical installation in June 1993.Installation of the rooftop array was completed on July 1 and 2, 1993. System electrical workwas completed on July 15, 1993. Startup and fictional check out were completed on July 16,1993.

Conceptual Design and Operation

The conceptual design of the original system which is illustrated in Figure 1, shows its mostimportant components -- the PV arrays, DC interfkce, batteryliiverter module and controlcomputer.

The system at NDGO was designed for summer peak shaving using the combined output of thePV array and energy in short-term battery storage. Based on data taken during the Summer of1992 in tests at the University of Delaware, and data collected during system testing from Julythrough September 1993, clouds are practically non-existent on the summer days in whichDelmarva experiences system peaks. On days like this, the output of the PV array is evenlydivided between morning and afternoo~ peaking at solar noon (approximately 1:00 PM EDT).However, Delmarva’s need for capacity, like many other utilities, is skewed towards the afternoonhours. To overcome this probleq the array’s output during the morning hours was used tocharge a battery bank. This stored energy, minus storage losses, was combined with the array’soutput in the aftemoo~ when it is most need~ and fd into the building’s distribution grid. Thesystem’s capacity is dependent on the amount of energy stored prior to dispatch and the durationof the dispatch period.

Octohw18ss Pege2

De&newePowercLbhtcompeny PV:BDNUSPhase1Fk@ RepoftCoopmdw AgrwnmntNo.DE-FC3643CMOS69

DISPATCHABLE PV SYSTEMBlock Diagram

UmDC/DC

m m -mm-b ‘c/Ac

mm

ConverterBatteries(26 KWH) Inverter

A

Rooftop PV Array Battery/lnverter Module

(6 to 12 KW) ( 111

480 VAC, 3 phase - d

I v

outputTransformer

Delmarva Power & Light Company

FIGURE 1

mhmuv8Fowu& ughtcampsny PV:BONUS Ftmsa1FinalRqad~ AgMelnwltk DE-FC3M3CHM669

VDC. The subarrays were originally series-comected with a center tap neutral at the inputterminals of the battery/inverter module to achieve the necesszuy input voltage.

The modules used in the array were manufactured by AstroPower. The modules are rated at 61Watts each at PTC and consist of single-crystal silicon PV cells encapsulated in EVA (ethyl vinylacetate). A glass-aluminum frame is used for mechanical protection. A total of 240 modules areused in the NDGO array.

The arrays are mounted on ballast~ roof jacks. This mounting system is unique because it doesnot require roof penetrations or additional structural steel. Each roof jack is bolted to agalvanized steel tray. The first step in the installation of the roof jacks was the removal of existingroof ballast. An additional layer of rubber roofing material was placed inside the cleared are% andthe ballast trays were placed in their proper locations. Once in place, the ballast was restored toeach tray and the PV panels were mounted on the roof jacks. The ballasted roof jack system issuitable for existing open roof areas where an additional dead load of 7 to 10 pounds per squarefoot can be supported by the roof structure.

Battery/lnverter Module

The batteryhverter module manages the energy produced by the PV amay by operating in one ofthree primary modes -- battery charging peak shaving or simple grid-connected operation. Theunit originally installed at NDGO was adapted for this purpose from an AC Battery Module,manufmred by the AC Battery Corporation of East Troy, Wisconsin. The AC Battery modulewas originally designed for utility load leveling and peak shaving using batteries alone. In itsoriginal configuratio~ the batteries are charged using off-peak energy from the utility system anddischarged to help manage short-duration peaks at substations. The AC Batte~ Module wasdeveloped jointly by the Omnion Power Engineering Corporatio~ the Delco Systems Division ofGeneraI Motors and Sandia National Laboratory. For the purpose of using an AC BatteryModule on the dispatchable PV systeW the necessary modifications included addition of a DCinput port to permit batte~ charging horn the PV array, and changes to the system controlsoftware.

The module is designed to provide nominally 25 kwh of energy storage when discharged overapproximately 4 hours. Conversion from DC to AC is accomplished using a 31 kW solid stateinverter, which also permits the use of the unit for reactive power (VAR) compensation. Becauseof the developmental nature of the PV charging software, the grid charging and static VARcompensation features were not incorporated.

With the configuration represented in Figure 1, the PV array operating voltage was determined bythe battexy float voltage, which was set at 621 VDC. The system output is 3 phase, 60 Hertz at

480 VAC. The inverter output is matched to the building’s electrical distribution system througha step up transformer. The inverter incorporates standard overcurrent, undervoltage andfrequency protection. The inverter will also isolate itselfin the event of a loss of external power.This will prevent “backfkeding” the grid during utility outages.

Octubw79s6 Page4

oe&n8w8Pmvwa Lightcampany PV:B~~ p- f Fitd xCoopwatlveAgreementNo.DE-FC36.93CMOS69

The original batieries used for this application were 48 Delco 2000 twelve Volt heavy-dutytruckhnarine batteries. The batteries incorporate conventional maintenance-i%ee,lead-acidtechnology. Although not trrdtionslly used in PV or stationary power applications, thesebatteries were chosen because they are mass-manufactured. Since they are also packaged into aneasily replaced module, a favorable trade-dfresults between initial capital costs and batteryreplacement costs.

Data Collection and Analysis

A significant amount of data is being collected on ambient conditions, PV system pefioimanceand building system performance. Ambient data includes insolatio~ temperature, humidity, andwind speed and dtiection. PV system performance parameters include array voltage, current, andtemperature, as well as the system’s AC power output. Building system performance parametersinclude the loads on the four primary f~ers, engine generators, four chillers, and four HVACmotor control centers. All of this data is logged at 15 minute intervals.

The data collection equipment and software were provided by &ension Technology. Data isdownloaded via telephone line to a PC for monitoring and analysis.

The data analysis presented in this report fmuses on component performance and resourceassessment. Specifically, the data is being analyzed to:

1. Evaluate the PV array performance relative to ambient temperature and insolatio~

2. Evaluate batte~ storage capacity vs. discharge rate;

3. Evaluate system losses and inefficiencies; and,

4. Compare Delmarva summer group loads to insolation.

Ocio&r f996 Page6

Dehanm Pimw&l.@tconlpanyCoopemtfveAgreement

PV:BOWS Pheee1Finall?eporlMO.DE-Fc36-63cnlo669

PROJECT COST SUMMARY

Capital Budget

A total project capital budget of $275,000 was established in Delmarva Power CBID Number 93-D-93. In additioxL$55,000 of R&D fimds were used to support the development of the hardwareand software for PV batte~ charging and the data acquisition system. A breakdown of projectcapital expenditures is shown in Table 1.

Table 1

NDGO PHOTOVOLTAICS PROJECT CAPITAL COST COMPARISON

CONTRACTOIUVENDOR ITEM COST

AstroPowe# PV Arrays and Balance of $186,653System Equipment

WSMW Roofiop Installation 7,500

Tri-M Electrical Construction 18,083Bruce Industrial Access Stairs and Roof Shelter 25,444

Miscellaneous Materials 423

Total Owner’s Costs Payroll, Loa&ngs and 32,432Miscellaneous

PROJECT TOTAL $270,535

Because the NDGO dispatchable PV peak shaving system is the first of its kind, caution must beused in comparing its capital costs to other PV systems. it is dfficdt to compare the capital coststo other PV systems. The installed costs of grid-connected PV systems without storage rangewidely ikom about $6,500/kW to $11,000/kW and higher, excludlng owner’s costs, special accessand data acquisition. The upper end of this range represents custom-designed, roof-mountedsystems.3 When owner’s costs, access stairs, and data acquisition system electrical installationcosts are excluded from the NDGO Project, the total capital cost is reduced to $208,000(approximately $14,500/kW). If battery storage is eliminated, the cost is reduced toapproximately $157,000, or about $10,800/kW.

For compariso~ it is more expensive to increase the size of the array in lieu of using somestorage. Since Delmarva’s peak loads occur at about 5PM during the summer, a fixed plate PVarray w“thout storage could only be credked with about half of its nameplate capacity. Ifa 15 kW

2AatmPw@scontmctindudedthepKcha8eofebctmaldaaign~andrniecellanaoua tlardwarefrOmAacamOnTechndqy,endthepumhaeeoftheACBetteryModuk.

3Recmtmeaapmhaaeaofernell,roofmountadPveyetameformekHlai“ appkethahave beenreportd atappmamatdy

s6,6WkW. Larger,“one-d-a-kid’ eyetemaare Conddembiymoreexpenam beceueeof increeeedengineeringand inetalletbnOoeta.

Octoberf996 Page6

DelmamlPoww&ughrcompaly PV:BONLJsPhase1Find RepoHCoqmmtiveAgreementNo.DE-FC36-93W-$

credited capacity is required over a three hour period from 3:00 PM to 6:00 PM a PV array of atleast 30 kW would be required, which would cost in excess of $300,000. This should becompared to the installed hardware cost of the system incki!ing storage of about $208,000. Forpeaks occurring later in the day, the cost of a PV system without storage would be even higher toassure a specified capacity level. While tracking could offset some of the added amay cost,significant additional costs would be incurred in modifications to the roof structure necessary toaccommodate tracking hardware and additional wind loadings.

PERFORMANCE OF THE ORIGINAL SYSTEM

PV Arrays

Immediately following initial installatio~ several modules were replaced. One failed when theglass on the module shattered, probably due to improper handling or fhme misalignment. Threemodules were replaced after measurements detected open-circuit voltage levels belowspecifications.

The AstroPower Model AP-5107 modules used in the PV array are nominally rated at 61 WattsDC at PTC. The output of the entire array, which consists of 240 modules, should be 14,500Watts DC at PTC, corresponding to an array “maximum power operating point” at 696 Volts DCand 20.8 Amps. In the original conflguratio~ the array operating voltage was determined by thebattery bus voltage. The actual operating point at PTC was therefore 621 Volts andapproximately 22 Amps, or 13,700 Watts DC. AstroPower’s performance warranty states thatthe DC power output will be within 10 percent of the predkted output. For a predkted output of13,700 Watts, the minimum guaranteed output is 12,330 Watts. Actual array output at theoperating point is approximately 11,600 Watts, or about 6 percent below the minimumguaranteed level.

The array output deficiency is probably caused by several modules in the east subarray which aresuspected of pefiorming below specifications. AstroPower has acknowledged that a deficiencyexists, and has resolved the warranty claim.

octctbwf996 Pege7

D2hmrvaRmw&ughtamPaJly PV:BOMB Phase 1 FinalRepodCoopemUveAvenf No.DE-FC38-S3CHI066$

DISPATCHABLE PV SYSTEMNominal PV Array Performance

,4DC KWPTC Quumltoa Point

12 —..-.-.-.. -—.-.——. ——.-.—-. —. .— -—.—.——..—-

10 - —.--....—--.-”-Prodietod MinhunI

8 -------------------------

c — ...—..-.--. —-... — .——. —- . . ...--. -...——

4t--

-.-.: -.-. - ......"--------------------------------------------- 1.

400 moo

Plane of Array insolation’$~/aq. meter)eoo

FIGUI?X 2

Fi~re 2 shows the predicted pefiormance of the PV array compared to actual performance at 20degrees Celsius for varying insolation levels. The line which represents actual performance at 20degrees Celsius is based on a regression of measured performance data.

AC Battery Module

The two most important pefiorrnance parameters for the AC Battery Module are inverterefficiency and energy storage capacity. In additio~ overall inverter and control system reliabilityand power quality are also discussed briefly.

The inverter, which is a modfied version of Omnion’s 3200 Series PV inverter, has petiormedwell since start up. This is significant because inverters have historically been troublesome ingrid-connected PV systems. The inverter’s output has been tested across its operating range,with no observed problems even at maximum output. Although the inverter is designed andmanufactured to comply with IEEE 519 for current and voltage harmonic distortio~ nomeasurements have been taken to date to veri& its performance in the field. Measurements takenin the flwtory indkate that the unit is in compliance with the IEEE standards in effect at the timeof manufacture.

The guaranteed inverter efficiency point is 93 percent at its rated outputof31.25 kW. Thisefficiency is determined by measuring the AC output power upstream of the system step-uptransformer and dividing by the DC input power. Inverter efficiency is a finction of output, andmeasured efficiencies show that the inverter operates at 93 percent or higher beginning atapproximately 12 kW. Efficiency from 6 to 12 kW increases nearly linearly from about 90 to 93percent. Efficiencies increase logarithmically from approximately 60 percent at 1 kW to 90percent at 6 kW. In the original design a significant inefficiency was introduced by the step up

~ 1996 Pege8

neIlwv8Plnwr&Lbhtcomp8nY PV:BONUSPhaee1Find RepiniCooperelhmAgreementNo.DE-FC36-9SCHI0669

trantiormer. The idling losses incurred by this transformer are 2 percent of its rated capacity,which corresponds to a 620 Watt continuous loss whenever the unit is connected to the grid.This problem was addressed with later modifications.

The amount of energy delivered from battery storage to the inverter of the AC Battery Module isa function of battery dkharge rate. Table 2 shows the expected module storage capacity for fulldischarges of 1,4 and 8 hours duration.

Table 2

Expected Battery Storage Capacity

Time Energy from Storage Energy from Storage Discharge RateMinutes DC kwh DC Amp-hours DC hll)S *

60 20 35 35.4240 25 44 11.1480 30 53 6.6

Figure 3 shows the measured battery capacity for twenty fill battery discharges from July throughOctober. Many more partial discharges were done during the same period, but do not provideinformation about total battery capacity. When comparing ener~ delivered from storage versusdischarge time the overall trend is a characteristic logarithmic curve. The correspondingdischarge rates are also shown. There are two explanations for the significant amount of “scatter”in the data. First, most of the points shown in Figure 3 were derived from data collected by thedata acquisition system at 15 minute intervals, which results in relatively poor resolution. Second,15 of the data points represent system dispatches, which included energy from the PV arraycombined with the battery. Because these discharges were at fixed output levels with decliningPV input they are characterized by large changes in battery discharge rate.

Octobw 19sS Page9

Ddm8rv8Pobvu&LightcoJnpeny PV:BOWS Phase1FimIwCOOPWWWA-ant Mo.DE-FC36-S3CM0569

DISPATCHABLE PV SYSTEMOverall Battery Performance

Amp-Hours (DC) Amps (DC). _60

60

40

30

20

10

0

+

-------------------------------------------------------------** J-“”-–”–-”-”--””-50.

. ,. .-. --—-—----—-9.- .. ..- .—. —. —.-. —

~:

——.—.--..————— .-

Amp-Hour8 of Storage40. . .

●.-—.— . . .. .. .... .. . . ... . .. . . . . . . . .. .. ...—. ——.—-... 30

+;9”.+ ..-—-.-.—....-.

+-----...-.."------.-..-"-"---""""---.---"`"--"-"-""---"--"-"`""--20~++

+++ +

..................... ............ ..... ........ . ... .... . .......... ..... . .. ............................. .... .............................&pa

10Average Dischar

o 100 200 300 400 500Time (Minutes)

FIGURE 3

To provide a basis for comparison five “controlled discharges” using only energy stored in thebatteries were done to exclude the influence of the PV array, and to more precisely measure theelapsed time to fill discharge and total energy released horn storage. The trend for thesedischarges is shown in Figure 4, and compared to the expected capacity cwve and the overallcapacity when PV energy is included.

When expected capacityis compared to measured capacity in controlled discharges, the data trendfor the amount of energy actually available from storage indicates that about 20 percent lessenergy is recovered from storage than expected for sho~ high power discharges. For longerperiods, however, the amount of energy recovered from storage was closer to what is expecte&although still about 10 percent less. Battery capacity data points for discharges w“th PV input arealso shown for comparison.

October 19% Page 10

DafmMvaPoweraLightcompanYCooperWveAgtwmentNo. DE.FC36-9SCM0669

PV:BCWUS Ptmsa1 Fti ~

DISPATCHABLE PV SYSTEMExpected and Actual Storage Capacity

Amp-Hours (DC)‘“~

1- x

50 --"-"-"--"---"""""""--""""-"-"-"--"""""---"n-""Y---`--"'-'`------- ....—---. ..........——-... ........—.I

L/ “ “/’ :,,..-/..*40 -“-’”””---”-”----””---”-;---

I

..--- v.=---- ...----------------------------------------------------,..,,---X

,------------ “

30“....—.-.a-+: -...--- ....--&------------------------------------------~., ●

>/=””

tF’””2“ ..-&------------------------------ ------------------------------------------------------

{

10t

... ......................................................... .................... .. .........................................

1o’ 1 t ! { I

o 100 200 300 400 500Time (Minutes)

— Overall Capacity Expected Capacity — Controlled Diooharge

FIGURE 4

The AC Battery control system accounted for most of the dficulties encountered during the 1993test period. The system control algorithms for charging dispatch and float operating modesreside in software on a PC-based controller. This controller was “off board” the AC Batterymodule. A separate controller in the inverter was used primarily for regulating the DC busvol~ge. This voltage level is critical for proper charge and discharge control. Several minorcommunications problems were observed in which the PC lost the abiity to communicate over theRS-232 link with the control board in the inverter. These were easily reset with no ill effects.

Voltage regulation on the DC bus was also a problem. Because an RS-232 serial communicationsport was used, the speed of communications was relatively slow. This made it diflicult to stabiiethe DC bus voltage, especially when the batteries were &lIy charged and the bulk of the PVarray’s energy was exported to the grid. This caused an oscillation in the DC bus voltage andrapid system power output fluctuations under these conditions, although there were no adverseeff’s on the system hardware. The last problem with the control system was an unexplainedsoftware error which occasionally caused the system to shut down. An error message was alwayspresent although the root cause was not determined.

At least part of the storage shortfhll may have been due to the charge control software. ACBattery believes that the control software may have prematurely terminated the batte~ charge

Ociwmr1s% Page11

De&rmwePower&ughtcomp2ny PV:BONUSPhaee f F&refRwCqwathre AgreementNo.DE-Fc3S-93CMOSS9

6ycle, resulting in battery undercharging. Other causes of the shortfall maybe due to battery cellreversal, or other battery-related problems.

Although the control and software problems were troublesome they did not prevent the systemfrom tinctioning for test purposes. Future generations of this system will have “on board”controllers, which should eliminate communications and voltage regulation problems.

System Dispatch Performance,.

The primary purpose of the NDGO system was to test the practicality of dispatchable PV systems.The system was dispatched a total of 27 times horn July through late September 1993. Anexample of a “back-to-back” dispatch is shown in Figure 5. This represents the most dficultdesi~ condition for the unit, because it requires the batteries to be filly charged in time for thepeak on successive peak days. In the mid-Atlantic regio~ system peaks usually occur in the lateafternoon. Therefore, battery charging takes place during the morning and early afternoon hours.The ability to be reliably dispatched on successive days is a fimction of the availabdity of solarenergy for battery charging on system peak days. This is explained in greater depth in the sectionon system sizing.

“BACK-TO-BACK” DISPATCHJU!Y 22 & 231993

I

Group Load

2000 -

... . ...... ..... .... ... . .... ........

eattorioa Chargad1000—

“%~ System Dlepatch --. ——... —---

Batt.rbc Clmrglng600 -.—----------- ........-...-..— - ....-.-.--.—...------- .. ........... —..,._,..-. .. ... ........

o93s7elt tst6t7nr21s8 la67ettt3mt7m212s

Time of Day

FIGURE 5

The pdormance of the system during dispatches and charging cycles helped to highlight a few ofthe problems associated with hardware components (i.e., control syst~ battery storage capacity,and PV array pdormance), but also established the soundness of the basic concept. This is a

02tober*9% Pege *2

De&M/w Powu 6 L&m company PV:BONU3Phese1FfnalRepott~A9 rewnentNo.OE-FC3M3CHM669

major step forward, simply because this type of system was never actually tested before theSummer of 1993. Refinements will be necessary before a commercially viable system is available.

Octubef1996 Page13

~R-fcMw-pyCOW-* Awuem@UIA DE-FCX-S3CMOS69

PV:BONUSphase1FtnelRepa#i

Tasks 1 & 2- Conceptual Design Refinement

Tasks 1 and 2 of the original Statement of Work were identified respectively as SystemConceptual Design and Building Conceptual Desi~ and were intended to provide the basis forsystem refinements in Phase 2. This was accomplished through the experience accumulatedduring the desi~ installation and operation of the test and demonstration system. Several of theearliest proposed refinements were later included in a Phase 1 contract modifkation.Tasks 1 and 2 were originally conceived as stand-alone activities. As work progressed, it becameapparent that these two tasks were more closely tied together.in Tasks 1 and 2 are therefore combined in this report.

FUNDAMENTAL DESIGN GOALS

The results of the work periiormed

A relatively simple system design was used to demonstrate the conceptual and technical faibilityof dispatchable PV systems. While this simple design was adequate for the purpose ofdemonstration and data collectio~ it was not intended to represent a refined, commerciallyavailable system. Before a reiined design could be completed, a set of fundamental design goalshad to be established. These goals reflect a combination of the designers’ experiences and marketpreferences (see Task 3, Market Research).

1. Modularity: The design of a dispatchable PV system should be modular. Modularityapplies to the design of sub-system components, as well as to the modular nature of theintegrated “units.” The use of modular sub-system components permits rapid design turn-around, and simplifies component sourcing assembly, installation and maintenance. Amodular “unit” design permits flexibtity in sizing systems (consisting of one or more units)for particular applications. Modularity also emphasizes the use of a controlled, factoryassembly environment for sub-system components, and minimizes the amount of fieldlabor required to install systems.

2. Functional F/m”bilitv: A variety of fimctions should also be inherent in the design of adispatchable PV system. Functional flexibility is required to accommodate diversecustomer requirements, and is already a common fare of many conventional buildingsystems,ranging from HVAC to emergency back-up systems. A key advantage offictional tlexibtity is that it adds value to the system. It is not a requirement that everydispatchable PV system be capable of performing every envisioned fimction, but that itshould at least be easily modtied during or after assembly.

3. Desire Flm”bilitv: A the market for PV systems evolves, it is certain to become moresophisticated and demanding. In additio~ new design standards will evolve which willrequire changes. The designers’ responses to new market demands and design standardsmust be timely, othenvise product and system offerings will simply be uncompetitive. Asystem with high design flexibility permits rapid responses to new demands without theneed for complete re-design.

October1996 Page14

Do&Mrv8Pobwr&LfghfcoinpulyQwpwath AgrwmentNo.DE-FC36-93CHMS69

PV:BONm Phase1FinalRepcut

The combined features of modularity, fictional flexibfity and design flexibtity will ultimatelyhelp make dispatchable PV systems more economically competitive. By emphasizing modularityand flexibility in products for the early marke$ system designs and fatures will not be fixedprematurely. Another characteristic of an evolving PV market will likely be the rapid introductionof new components, processes and applications. Design flexibtity in an older product or systemmay not help it compete against newcomers, but it helps foster the institutional flexibilitynecessary to smooth the transition between “old” and “new technologies.

SYSTEM DESIGN ISSUES

With the above goals in mind, many system design issues were identified during the initial desi~installation and operation of Delnxuva’s dispatchable PV system. Design issues are categorizedby sub-system and are subsequently addressed in the Phase 1 contract modification and/or Phase2 planning.

PV Array

The PV array is the primary source of energy for the dispatchable PV system. The widespreadintroduction of PV arrays to the commercial building market must account for the requirementsbuilding owners and tenants, code officials, and those responsible for facility maintenance andoperation.

of

1. UL Listing One of the main criteria used by code officials in determining the suitabilityof electrical equipment for a given application is a listing by Underwriters’ Laboratories.Presently, many manufacturers of PV modules do not list their products with the UL.While some code officials may be flexible regarding the requirement for UL listings in“experimental” systems, it is preferable in the long-term to have PV modules listed withthe UL to minimize Iiabfity and code compliance problems.

2. Arrav Oven Circuit thlfa~e: The open circuit voltage of the PV array determines theNational Electric Code voltage classifmation of downstream equipment as well as systemelectrical design. The threshold between “low” voltage and “high” voltage is nominally600 Volts. The array open circuit voltage level encompasses saflety,design andconstruction issues. The demonstration system’s array open circuit voltage, which wasdetermined by the input voltage necessmy for the AC Batte~ Module, was approximately880 Volts DC. Electrical equipment classified for high vokage DC semice is nearlyunavailable. This problem was addressed by using a center-tapped array contlguration inwhich two separate array source circuits were installed operating at nominally 440 VoltsDC. While this solved the immediate problem it required redundant cable, conduits,disconnects, junction and pull boxes.

3. Lirhtninz Rotection: Lightning protection is required on many commercial buildings.The design and installation of lightning protection systems must be done by certifiedlightning protection contractors. In many cases, especially retrofits, simply connecting the

Octotnu1s96 Page16

Da41uw8Fawu?4ughtcompany PV:B~USPhaso 1Fin#Rap-iCoopemiiveAgmamenlM. DE-PC.36-93@W669

PV array to an existing lightning protection grid is not sufficient. Although thedemonstration system did include lightning protectio~ different requirements may apply toalternative structures.

4. fhnaZZvs. Lame-Area Modules: The demonstration system utiliied AstroPower AP-5 107modules. The dimensions of these modules are 21 inches wide by 46 inches long. Sixmodules are assembled into a panelized assemblies on aluminum channels. Ultimately,panelized assemblies of small PV modules will be inefficient. Each assembly requiresintermnnecting wiring, junction boxes and structural supports To the extent that large-area modules can be substituted for panelized modules, fkto~ assembly costs should bereduced. The elimination of parts, especially electrical comectors, should also enhancearray reliability and reduce field maintenance and troubleshooting costs.

5. PVArrav Electi”cal Clmneti”ons: Individual PV modules were electrically connected inthe factory prior to shipment. Once at the site, panelized assemblies were interconnectedusing “quick” connectors, and finally terminated in row junction boxes to create sub-arraysource circuits. The row junction boxes were then interconnected for each sub-array. Thedemonstration system required the electrical sub-contractor to install a considerableamount of the interconnecting cable and conduit. In part, the arrangement of the electricalconnections was dictated by the use of a ballasted roof mounting system. For fitureinstallations, the number of electrical interconnections requiring the use of a fieldcontractor should be minimized.

6. Buildin~Zntezrated PVi%tm”ak: Although the Phase 1 development workconcentrated on the use of crystalline silicon PV arrays, thin-film PV arrays, or PV roofingand glazing can also be used in dispatchable systems. The main issue, in this case, is thestandardization of PV source circuit designs.

PV Array Support Structures

Many engineering installation and operational issues are encountered in the design of PV arraysupport structures. Additional complications arise when the ages and types of buildings androofing systems are considered. The Phase 1 work concentrated on support structure options forPV arrays mounted on low-rise, flat-roof commercial buildings. Table 3 summarizes the mainissues.

Dctob# 1996 Page16

TABLE 3

SUMMARY OF SIGNIFICANT ISSUES - ROOFTOP PV INSTALLATION METHODS

Caragoty

Efw~n

opemtfo@hfainteMnca

New Structure Retrotlt StNctura 1

Arohitectaand bWderaarer@ Thaoqacitytoreaiat addiil I

Uaedtoauowingfwthak@ationof Pvaupportamduras.Codaa,standardsSndpmauraaarendwaudefined.coordinationwitharchitectand overallprojaottiming are oritical.

Windbadingforoeeandmorrr@aon&dldingatruot-mwnbsramaylirnitthaPVafraysize.

irwfaranceaWilhotharfooftopq- (HVAC,etc.)maka

ISnay tayoutdimoult.

Rooftop paneHha willprobably be required, addingOoattothaovaraff inatalfation.

F* Structural modif~

I may berequired] Roofwanantymaybe altarad

aa a resultd-making---

AratWtdrudura mayaffeotbuildii seathab andarohiiactml integrity.

Ballasted Roof Jack

Addiil bdiaat tW@Sdtoraaiatupfift mayaxoaadtha

~ deSgn limitsof theroof.

,.

Interferences with other rooftop~~t (WAC, dC.) IIWy

maka Iaput difficult.

Roottop drainage rsquiramankmayraquira aksathanoptknallayout.

Saismiidesign criteriamaypraciudethauaeof bauaatadsystem in palta of the COuntly.T~ or othef faatankqmathodamayberaq uired.

PVayatammuatbaremovadfor roof replacement or rapaira.

Roof warra nty maybe altered.

AC Batte~ Module

A primary goal of the PV:BONUS product development effort is to reduce design and fieldinstallation costs by incorporating f~ory-produced, standardized sub-systems. Although theoriginal design of the AC Battery Module was not intended for use in a dispatchable PV systemone of its chief advantages is its modularity. The unit as it was originally configured had severalshortcomings. These were:

1. (krti”np Vokwe The operating voltage of the PV system had to be in excess of 600VDC in order to charge the batteries. IX electrical equipment suitable for this voltagelevel is virtually unavailable. This requires a bipolar array design with duplicatedisconnects, cable and conduit, resuking in higher equipment and field installation costs.

2. DCZn@%ace A DC contactor was used to connect the PV array to the AC BatteryModule. This forces the PV array to operate at the battery bus voltage instead of the PV

0cto6er1S96 page17

Debnarw-or& Ught company FV:BONUSPhase1Find RqnntCa@wrativeAgrewnentNG DE-Fc36-93CNMSS9

arraymaximum power point. The addition of a DC to DC converter will permit maximumpower tracking and allow the array operating voltage to be reduced.

Controik: The AC Battery Module was controlled by a stand-alone desktop computer.3.While this was acceptable for test purposes, it is not viable in production units located incommercial buildings.

4. PCS Om’im”zation: At this time, the optimum comb~tion of equipment for a production“unit” includes a PV array between 6 and 12 kW at PTC (depending on designconditions), 25 kwh of battery energy storage and a 15 kW PCS. The current PCS isoptimized for 32 kW. A DC to DC wnverter will permit the PCS to operate at itsoptimum vokage (600 VDC) while halving the PCS current. This will result in reducedPCS production costs.

5. IZlm”ble Charm”ng: The unit is presently not capable of being charged from the utilitygrid. The ability to charge the batteries from either the PV array or the utility grid willhelp optimize battery life and enhance flexibility.

Contract Modification to Address Early Design Jssues

Delrnarva Power’s original PVBONUS proposal included the installation and testing of adispatchable PV peak shaving system. The intent of this task was to prove the technical feasibilityof dispatchable PV systems and to identfi areas requiring additional technical development. As aresult of testing throughout the second half of 1993, two major development areas wereidentified. The first area of development is the refinement of the Batte@Inverter Module for thedispatchable PV system. The second area is in the development and testing of cost-effective roofmounting structures. The Department of Energy requested an expansion in Delmarva’s scope ofwork to address improvements to the batteryliiverter module. Work on roof structures will beincorporated into Phase 2..

In order to address the specific technical issues associated with the AC Battery Module, the unitwas dkccmnected and returned to the manufacturer in March 1994. AC Batte~ incorporatedfeatures which considerably improve its performance and enhance its commercial viabiity. At theend of the Phase 2 development and testing effort, only minor modifications should be necessaryto field commercial prototypes.

PERFORMANCE OF THE MODIFIED SYSTEM

In order to address the shortcomings of the original system severakmodifications were completedas part of a contract amendment.

Octohw1986 Page18

Ddkn8w8Pawu&ugM conlp8ny PV:BOWS P- f - WCOOpmtflve AgrwnmntNo.DE-PC36-S3CHWS6V

PV Arrays

The high operating voltage of the PV array, which originally operated in excess of 600 Volts DC,was addressed by a reconfiguration. Originally, the two subarrays were configured in series tocreate a bipolar, center-tapped array. This was done to accommodate the high operating voltagenecessary to properly charge the battery/inverter module. Whh the reconfiguration of thebatterylkmrter module, high array operating voltages were no longer necessary. The subamayswere reconnected in parallel, resulting in a monopolar array configuration. The operating voltageof the array is now less than 600 Volts DC.

The array reconfiguration and reduced array operating voltage was made possible by the additionof a DC to DC converter to the batteryhnverter module. By adding the DC to DC converter, it is”also now possible to employ PV array maximum power tracking. Based on data taken shortlysfler the system modifications in December 1994, PV array pefiormance improved by about 10percent as a result of maximum power tracking. This is shown in Figures 6 and 7, which showregression lines for the PV array output as a fimction of Plane of Way Insolation for December1993 and December 1994. While this improvement is representative of pefiormance during thewinter months, when the array tends to operate at lower temperatures and higher voltage levels,the overall annual improvement is expected to be about 5 p~cent.

—

ARRAY OUTPUTWITHOUT MAX POWER TRACKING

December 1993

10-1 4

0 200 400 600 Soo

POAinsolation (VUalts/Sq.Meterj

FIGURE 6

DdmanmPowefcfJ@trconIpany PV:BONUSPhase1Find ReportCoopemfhmA~ M. DE-FC36-SS~

ARRAY OUTPUTWITH MAX POWER TRACKING

December 1994

1

0o m m m Soo

POA Insolation (UVatts/Sq.Meter)

FIGURE 7

AC Battery Module

Significant changes were made to the battery/iiverter module by the AC Battery Corporation.The first significant alteration was the addition of the DC-to-DC converter. The converter servesas the interface between the PV array and the battexyhmrter module. As previously described,the converter permits the PV array to operate below 600 Volts DC and permits maximum powertracking while simultaneously allowing the battery string voltage to be maintained at optimallevels.

The original control hardware configuration was also completely changed. The stand-alone PC-based control system was replaced by an on-board control computer with a keypad user interface.These improvements move significantly towards a commercially viable system by making the unitmore self-contained and “user friendly.” In addhion to changing control hardware, controlsoftware was also changed. The new software, developed to optimize the operation of thebatte@iiverter module with PV, has exhibited none of the problems and errors associated withthe earlier software versions.

The combined modtications of the battery/iiverter module have added several capabilities,including the ability to charge the batteries from the PV array or grid, automatic “wake-up” andshutdowq remote dispatch control via a building energy management or SCADA systeW and theability to pre-set charge/dispatch schedules.

Defmuve Power&ugMcompany PV:BOWsFhese 1FhalRe+P@CooperativeAgm#nentNa DE-FC26-92CHW669

Overall DC to AC eonversion efficiency is slightly lower than before the modifications, primarilybecause of the addition of the DC-to-DC converter. Although the converter accounts for someadditional conversion losses, these are offket by the ability to operate the PV array at its maximumpower point. Figure 8 shows the DC-to-AC conversion curve for the modified system.

DC to AC Conversion Eticiency

lm , 1 I I I f

o~o 2 4 6 8 10

DC Input (kWV)

FIGURE 8

OUTPUT CAPACITY vs RUN llME

30,I I I i

p~

it-l \ I I I I

FIGURE 9

October 19s5 Pege21

ne4nawaPower&LigbtconJpmy W:BOMUSPheseqFJrd ReporiCooperathmAgruemenf No. DE-Fc3s-9=HI0669

System Dispatch Performance

Based on preliminary tests following system reconfiguratio~ the expected battery capacity isshown in Figure 9 for an 80 percent depth of discharge. This figure does not include thecontribution of energy ilom the PV array (as would normally be expected), and therefore showsthe minimum time the unit should run at a given output level.

RESOURCE AND SYSTEM SIZING ANALYSIS

Resource Availability

Solar resource availability on system peak days is an important f-or in the operation of thedispatchable PV system. Because the system uses storage, it is not critical that the solar resourceand load coincide exactly. In Figure 10, the coincidence between Group Load and GlobalHorizontal Irradiance is shown for the highest hourly loads during the month of July 1993.Although the coincidence is good for about the ten highest load hourg it drops quickly. This isbecause there are very few hours when peak loads and solar energy precisely coincide, and is dueto Dehnawa’s late afternoon peak. The main implication is that the effective load canyingcapability of a PV array without storage in this region is significantly lower than the nameplaterating for all but a few hours.

LOAD & SOLAR COINCIDENCEJuly 1993

Fraction of Monthly Peak1.21

I GrouD Load I1- . .. . .. ... ... ... . ....... ... . .... .. .. ........... .... .............. ............... .............

0.8

0.6 “.._w . . ..#.–-...- ----. -... — . . .. —.-—-...

*

0.4* *--. _-. _-_.. -_--.-_-".._.._.-_. - ...__________________ ++---—----”

*

t

0.2 --.-----.-.-.---.-.--.-.-.-....-.--.-....,------------------------------------------------Global Horiantal Irradiance I

0’ , , , I

o 60 100 160 200 260Hours

FIGURE 10

October1996 Page 22

Da&naw8Pobvu4 ughrcompany W:BON@ w 1FinalReportCoopmtW AWWMXW No. OE-FC36-V3CHI0669

Figure 11 shows the coincidence between TotuZDaiZySolar Energy Avm”hzbility(expressed askWh/day/sq. meter) and DuiZyPeuk Loads for weekdays from July through September 1993.Because the trends shown are nearly parallel, this figure indicates that, for about the 20 highestsummer peak loads, there is a high and relatively predktable level of solar energy available on

those dizys. Because the hottest days in the region are also quite inuni~ the amount of solarenergy actually available is somewhat lower than clear, cooler, low humidity days. For lowerutility loads, the data indicates less predktabiity. This is because cooler summer days in the rnid-Atlantic region (also days in which utility loads are lower) area mixture of cloudy and very clearsunny days. This figure has a direct impact on the design of dispatchable PV systems becauie itindicates the relative quantity of solar energy available for collection and storage on peak days.

DAILY PEAK LOAD & SOLAR ENERGYJuly through September 1993

Fraction of Period Peak1.2/ I

I Daiiy Peak Load I1

O.a

0.6

0.4

.... ...... ..... ............. ..... ................................ ...... .... ....... ......................................-

~-~

.._..--._._...._.__..*____..... .... ....

.—-— *

-

..--—.—-—..._________.-—-. --------------._--. -_._.-.... .. ....,,..,.,,,_.Totai Daiiy Soiar Energy

~

I............................a.................2.....?..... .... .............. ..........?&..%..............-0.2* **

OJ 1 ! 1 t I Jo 10 20 30 40 50 60

Days

FIGURE 11

A similar analysis was completed in May 1995 to determine the coincidence between solar energyand Dehnarva Power feeder and substation loads. The outcome of the analysis indicates, ingeneraI, that a larger number of feeders are candidates for dispatchable PV systems than for non-dispatchable systems. This analysis will help quant@ the potential “distributed benefits” ofdispatchable PV systems.

October le# Page23

DehmMv8Paww&L4@~y PV:BONUSPhase1FinalRepo?iCOOpemweAgreementNo.DE-FC36-93WMS6V

System Sizing Methodology

The combination of component peflormance and resource availability dam along with reasonabledesign assumptions can be used to help size dispatchable PV peak shaving systems. Thefollowing steps describe the methodology.

1. Establish the number of times peak shaving is required during the peak period.

2. Using daily peak load andenergycoincidencetit% est~fish the ~aY “design~aY”.condhions, i.e., insolation values and ambient temperatures.

3. Calculate a PV array output curve representative of the “design day” conditions.

4. Set the time of day, duration and output level for the system.

5. Calculate the proportions of energy which must be supplied from the PV array andstorage.

6. Calculate the size of the storage block and PV array, accounting for storage round-tripenergy losses, and ambient temperature derating of the PV array.

7. Using battery storage curves and fixed array size, determine the dispatch output of thesystem for different times of day.

The above methodology is used in a simple spreadsheet analysis to predict the petiormance ofdispatchable PV systems, and the relationship between PV array siie and battery storage. Figure12 shows examples of the relationship for “design days” with insolation peaks at 650 and 1,000Watts/sq. meter. 650 Watts/ sq. meter is derived from analysis of the data shown in Figure 11. Ifit is determined that the system must be dispatched during the highest 20 peak loads, then theinsolation profile on the twentieth day will peak at about 650 Watts/sq. meter. This establishesthe insolation profile and the minimum amount of energy available to the system. For sizing thePV array properly, it is assumed that the ambient temperature on these days will be around 100~ (38 oC).

Figure 12 shows the relationship between the PV array size, dispatch capacity for a range of hoursand the available AC Battery storage capacity. In this figure, the utilization of the battery storageis maximized for each dispatch duratio~ with the size of the PV array determined by the energynecessary to filly charge the batteries prior to dispatch. In this example, all of the dispatches areassumed to end at 6:00 PM daylight time. Array size is nearly a linear iimction of dispatchduratio~ ranging from 4 kW to about 12 kW DC at PTC. Dispatch capacity decreases from 15kW to about 10 kW AC over the same range of dispatch duration.

October1s95 Pege24

DehelvaPowor& L#@rcompany PV:BONUS Phaee 1 Unel ReptWtrvemm!Mu DE-FC36-S3CHM6M~ Ag

PV ARRAY SIZE AND STORAGE CAPACITYFor Dispatch Periods Ending at 6:00 PM

so

storage Capacity (Amp-Hr8)40 -“-–---”-””-–--”-”-”-”””--”---”-””””””””-”---..... . ........ ............ . . ... .... ............ .-..--_._ .....—-----

___...—..-.—..— -----

-`-"-"-"--"-"-"-""""-"-"--""""""-"-"--"-""""""--"--""""-"--"-"-"""--65o-"-w7rn2--"--"--"""-"-"-"-"""---`20 Dispatch Capacity (kW, AC)

‘0 -“-”-”-”””--’~f~=~”c’c-0’ 1 I i I b 1 1

-o 1 2 6 7 6

Dispa;ch Dur~tion (~ours)

f?KWRE 12

The effect of a better insolation profile is to reduce the size of the PV array in relation to thedispatch capacity. For example, if the peak insolation is 1,000 Watts/sq. meter, the array sie inFigure 12 ranges horn less than 3 kW to about 8 kW DC at PTC. If this is used as a designcondhion in the mid-Atlantic regio~ it also reduces the number of peak loads for which thesystem can be dispatched.

Demand and Energy Impacts of the Dispatchable PV System

The technical impacts the dispatchable PV peak shaving system can be subdivided into energy anddimuznd components. The energy (kWh) output of the system results in a simple decrease in theelectricity purchased or generated from conventional sources. Because the system is alsodispatchable, it can be used to shave peaks, and effectively help to reduce the maximum demand(W) for power. The impact of peak shaving is dependent on the end user’s load shape. Endusers with relatively flat load profiles will experience less “flattening” or peak shaving than endusers exhibiting more pronounced peaks. The AEG market research helped to identbj categoriesof end users which have characteristics more favorable to the use of dispatchable PV systems.

odder f996 Page26

De/mnf8Poww&Lighf cantpany W:BON_ Phase1FhanlRqxxtrwnwntNo.DE-Fc36-93CHM69~Ag

The magnitudes of the energy and demand impacts are roughly proportional to the number ofmodular units necessary to achieve design objectives. The impacts of a single modular unitlocated in dii%erentareas of the country are shown in Table 4 for five cases. This table shows theapproximate annual energy output of the system (a fimction of average annual insolation),credhed demand reduction and the benefiticost ratios from both a utility and customerperspective, and for both dispatchable and non-dispatchable systems. The calculations for thesecases were done using a spreadsheet model under development by CEEP.

TABLE 4

IMPACTS OF A MODULAR DISPATCHABLE PV PEAK SHAVING UNIT ‘1)

Case Approximate Credited Demand Customer UtilityAnnual Energy Reduction (kW) Benefit/Cost Benefit/CostOutput (kWh) Ratio Ratio

Diapltdl- Nan- D&p9tdI- Dhpatch- Nan- D&patch-IJNZ able D&pstch- Sbk wNa able lMspatch- abk

8bk ●bk ●ble able

DelmarvaPower 16,800 16,800 4.3 14.9 0.63 0.76 0.48 ,0.66

SacramentoMunicipal 20,100 20,100 3.2 17.2 0.62 0.73 NIA N/AutilityDistrict

City of Austin 19,000 19,000 4.9 15.3 0.63 0.75 NiA NIA

Ecw (2) 16,400 16,400 4.4 12.2 0.74 0.98 0.51 “0.52

NiagaraMohawk ‘3) 15,300 15,300 0.7 4.5 0.61 0.69 0.40 0.41

Soutw Bynej d. al., W95,Photoval%aicsforDeman&SideManagement UtilityMarkets: A UtilityCustomer Partnership Approach

Calculating the economic impacts of the dispatchable PV system are complex because they aredependent on system ownership, tax treatmen$ utility rates and costs. A considerable amount ofwork in this area has been done by CEEP. Starting from the perspective of a utility customer, theprimary economic benefit derived from the use of a dispatchable PV system is a reduction indemand and energy charges. A utility customer would be expected to maximize benefits based onload shape and electric rates. Depending on the rate structure, the highest economic impact maynot coincide with the greatest peak reductions. For example, the owner of an office buildiig with

Octdnw19$6 pagem

DdmaMPow8f& ughtcompany PV:BWIJs Phasef FiIwIReporiCmqmmth AgnwnentNo.DE-FC36-93CHM66S

electric heat pumps may have higher demand in the winter than in the summer. However, higherdemand charges may result in higher b~s during the summer. The CEEP analysis shows thatdispatchable PV systems have a significant cost-benefit advantage over conventional non-dispatchable grid-connected PV systems, based solely on customer bfl savings.

From the perspective of a util~, the primaxy economic benefit is the revenue collected as a resultof the sale of capacity and energy from the utility system. The maximum benefit to the utility isachieved when the capacity and energy are available for sale at the time they are of highest value(usually at the time of system peak loads). In additio~ utilities can use dispatchable PV systemsas distributed generating units. In this role, the systems maybe able to defer a larger capitalinvestment. The CEEP analysis shows that the dispatchable PV system once again has asignificant cost-benefit advantage over conventional PV systems. The advantages for utilities arenot as great as they are for customers, primariiy because of tax benefits which cannot be claimedby utilities.

The CEEP analysis does not assume that there is good coincidence between commercial customerpeaks and utility peaks. In fact, because the customer’s peak demand is not necessarily coincidentwith the utility’s peak load, the utility may still be faced with the need to invest in incrementalgenerating transmission and/or distribution capacity, and still lose a portion of the customer’srevenue. Traditionally, the ownership of PV systems has been restricted to either the utility or thecustomer, although recent work by CEEP has also explored various utility-customer partnershiparrangements which help to overcome some of the economic and institutional barriers. Forexample, it is possible to create an arrangement between the utility and customer which allows theuse of the system by the utility for peak shaving purposes. In this way, both the customer and theutility can benefit horn the installation of the system.

PHASE 2 SYSTEM CONCEPTUAL DESIGN

The conceptual design of the dispatchable PV system is based on the experience gained in Phase1, particularly from the deployment and testing of the demonstration system. Basically, two typesof systems evolved from this work - small units for modest commercial applications and largepad-mounted systems for bigger commercial and utility applications. Supplemental marketresearch also revealed that off-grid versions of the dispatchable PV system could have greatpotential for village power and remote power systems. The features of the systems are describedin the following sections.

PV Array Design

The development of the dispatchable PV system is not restricted to the use of a singlemanufacturer’s PV modules. A disadvantage of this approach is that it makes standardization ofarrays and source circuits much more diilicult.

The standard PV array source circuit design must be based on a ceiling voltage level of 600 VDC(open circuit) at STC conditions. This level was chosen to stay within the design and safety

Odobw 7996 Page27

09h?Uw8Power&LighteonlpnyCoqwmtlve Agnwnunt No. DE-FC3643CUW69

W:BWm Phaso1F/d ~

requirements of the National Electric Code for systems at or below 600 Volts at Standard TestConditions. Based on manufacturers’ performance Normatio% “standard” monopolar PV stringarrangements are shown for several manufacturers in Table 5. The minimum amay size for adispatchable PV system is the same as the minimum -“rig size for any manufacturer’s modules.

TABLE 5

MONOPOLAR PV SOURCE-CIRCUIT STRING ARRANGEMENTS

Manufacturer Model Modules per Open circuit kW per Stringstring Voltage at STC

Solarex MSX-64 24 497 1.5

Siernens M-53 24 521 1.3

AstroPower AP-67 24 494 1.6ASE Americas GP-50 6 372 1.7

Smaller PV modules (SolareL Siemens, MroPower) would be panelized into larger sections.Large area modules will not have to be panelized prior to shipment. In some cases, the panelizedsections or large area modules can be mounted onto shop f~ricated structures and pre-wiredbefore shipment. The iinal configuration of pre-assembled arrays and structures will depend onshipping size limitations.

Support Structure Options and Configurations

The PV array support structure is an important component in the overall system design andinstallation. The design of the support structure is closely linked to the configuration of the PVarray. The design of cost-effkctive rooftop structures for PV arrays has been an elusive goalbecause of the lack of standard-ion and relatively limited experience. The diversity ofresidential and commercial rooftop materials and structural configurations, as well as buildingcode requirements, has only compounded the dficulties of developing cost-effixtive designs. Itis unreasonable to expect that a single design could be capable of satis~ng all structural desi~building code and performance criteria across the country. The approach taken by thedevelopment team is to offer flexible options which can satis~ a range of requirements, can beeasily modified and are inherently simple. Two options are available for systems deployed inPhase 2. These are described below.

1. BaZla@edRoof Jacks: Ascension Technology has developed and installed numerousballasted roof jack mounting systems. This system utilizes a series of galvanized steel“trays” filled with ballast to anchor the PV arrays. Each tray is fitted with a roof jackwhich is designed to accept panelized or large area modules. Acension Technology hasdeveloped array drawings and electrical interface drawings for SolareA AstroPower and

October19s6 Page28

~—-&-J~w#-w’w W:BONUS Plmse f FhmJReptutrvwmntNo.DE-FC36-S3CHMS@

ASE Americas PV modules. The ballasted roof jack mounting system is best suited toretrofit applications.

The ballasted roof jack mounting system can be used on flat roofs with the capability ofcarrying between 7 and 14 pounds per square foot of additional dead load.

2+ l%e-fdricatedSumort Structure: The alternative to the Ascension ballasted roof jacksystem is a pre-fabricated steel support structure, suitable for both retrofit and newconstruction applications. The PV array support structure will be a “monotube” structureconsisting of a single 30 foot section of 10 or 12 inch Schedule 40 steel pipe. Crostiswill be used to support the panelized PV modules. Each structure will support a total ofabout 3 kW of PV modules. The support structure would be used in cases where:

a. Ballasted supports cannot be used due to roof loading limitations, or local coderestrictions;

b. Other interferences require an elevated array,An array would be installed on a new building; or

: Anew roof is being installed.

The design of the proposed structure emerged from an investigation of steel structures for one tothree story, low-rise commercial buildings. These buildings are typically designed with columnbay spacings ranging from 20 to 30 feet. For arrays with moderate tilt angles (10 to 20 degreesabove horizontal) modifications to structural ilaming are usually not required if the loads aretransferred dwectly to the building columns. Based on this finding, each pre-fabricated structurecan be mounted to two simple stanchions welded to the buildiig coh.unns. There are currentlysome limitations on this design:

1. Generally, each structure must be installed over two separate columns, i.e., two PVsupport structures cannot share a common center column. This is because the centercolumn(s) will experience twice the loa&ng of the outboard columns. While suchinstallations are not necessarily exclud~ they must be analyzed case-by-case.

2. Because the structural attachments require cutting through the roof and welding,installation of the attachments must generally take place when the building is unoccupied.The potential for roof leaks exist% although this can be controlled through the use ofqualified contractors.

This structure utilizes low co% readily available steel components, and requires only simplef~rication techniques. A sketch of the proposed structure is shown in Figure 13. Because thestructure can be shop assembled and pre-wired, field labor is minimked. The design of thisstructure also lends itself to variations which can be used in parking lots and other open areas.

OdolKtf ~w6 Rage 29

Da&MwaAnuar&ugMcanpany~AUI—m—f No. DE.PC36-93CM06S9

PV:BONUS Wwaa 1FM Rapwt

PANELIZEDSUB-ASSEMBLIESARE PRE-VVIRED...

AND MOUNTEDTO A LOW-COSTSTEEL STRUCTURE

FIGURE 13

AC Battery Module Design and Functional Specifications

The AC Battery module used in the demonstration system required several modifications to makeit better suited to commercial applications. An improved version of the AC Battery module fordispatchable PV applications was completed under a contract amendment. Several new fatureswhich address the shortcomings of the original unit are described below.

1.

2.

3.

4.

The nominalAC Battery petiormance characteristics are 25 kwh of energy storage over a4 hour discharge period. These overall specifications will be retained, although improvedbatteries will be used (Delco AES 2010 in lieu of the original AES 2000). The unit’sinverter capacity is presently set at 32 kW. An inverter with less capacity maybe bettermatched with the battery and PV array, resulting in lower inverter costs. This will bedetermined later, once more operational data is available

Each AC Battery module will be configured with a DC-to-DC converter capable of up toabout 14 kW input. An option for a 7 kW DC-to-DC converter is under consideration.

The AC Battery module control system will include all protective and control systemfimctions, and can be remotely dispatched by either the utility or a building managementsystem. The operator interface will be a simple keypad located on the module instead ofthe microcomputer used originally.

The unit will also be capable of charging from either the PV array, the utility gri~ or acombination of both sources. This will assure that battery charging is alwaysaccomplished in a controlled manner, which will help extend batte~ life.

Octobar V89S Paga30

Debmuv8PobYw& L@htcampaIy W:BOMJ3 Wmsa1FitmlRepariCoqwnth AgnwnmntM. DE.FC36-93CHM669

System Packaging

The dispatchable PV system developed under this program will consist of modular “buildingblocks.” Atypical building bloclq shown schematically in Figure 14, includes a PV array whichcan range in size from 6 to 12 kW (depending on the local solar resource), and a batteryfinvertermodule. The batteryliiverter module includes 25 kwh of battery energy storage, a 32 kWinverter, and system controls. Because these units are modular, they can be combmed asnecessay to meet specified peak shaving requirements. For larger installations, building blockunits can be containerized. Each container is capable of storing up to 150 kwh of energy, with a

\ maximum output of 192 kW. Modularity also reduces the amount of field labor required forinstallation.

There are two basic units requiring installation - the PV array and the batterjdiiverter module. Ifa rooflop PV array installation is desire~ either the ballasted tray system or the pre-fabricatedstructure can be used.

Individual battery/inverter modules would normally be installed in a service room inside abuilding. The unit is virtually self-contained, requiring a small vent po~ and electricalconnections to the PV array and the building or utility electrical distribution system.Containerized systems would normally be provided as outdoor, pad-mounted units.Containerized systems include HVAC for temperature control. The typical system output voltageis 480 VAC, 3 phase, 60 I@ although other output specifications can be accommodated.Installation of a complete system is expected to take several days from the arrival of theequipment.

Because the system is modular, a wide range of output specifications can be met withoutdficuhy. One of the advantages of the containerized system is that it can be used to providepeak shaving service to individual buildings, or on utility distribution feeders. In this way, the

October19s6 Page31

system can be considered a distributed generation unit.

●V ARRAY /

m

FIGURE 14

/

/

DdmBrv8Powu& Lightcumpany~~ Air-m-h DE-FC3643ctUO-

W:BWU3 - f Wml Rqxsrl

Near-Term and Mid-Term Cost Estimates

Market assessment work performed by the Applied Energy Group, as well as work performed byCEEP and others have helped to establish the value of the dispatchable PV system under differentownership, electric rate and payback scenarios. As expected, there is a very wide range of valuesfor the system. Table 4, which summarizes the benefit/cost ratios for several case studies, showsthat in areas where electric rates are hi~ the system cost is near the break-even point already forutility customers. In areas with more moderate electric rates, the benefit/cost ratio is lower. Forbenefitkost ratios ranging from approximately 0.5 to 1.0, the implied system values range fromapproximately $50,000 to $100,000 per unit. hswning that 10 kW PV amays are used asproposed in Phase 2, this translates to a range of $5,000 to $10,000/kW. This range is importantbecause it helps to set the targets for the cost of producing dispatchable PV units for the domesticmarket. Table 6 presents cost estimates for fidly installed units now and in the near fbture. Theseestimates are, of course, sensitive to production levels, which can change significantly if units aresold overseas. In certain areas of the world, the cost of PV-generated electricity is alreadycompetitive with conventionally generated electricity. If a significant number of units are soldoutside of the domestic market the costof production could be much lower in the near-term.

TABLE 6

DISPATCHABLE PV SYSTEM COST ESTIMATES

Time Frame System Cost PV Module Cost Balance ofRange Range, PTC System Cost

Range

($llcw) ($lkw) ($ikw)Base (1994) 10,000 4,300 5,700

1996 8,500-9,500 3,900-4,200 4,600-5,3001997 7,000-8,000 3,200-3,600 3,800-4,4001998 5,500-6,500 2,500-3,000 3,000-3,500

The estimates above are basedon an internal study of PV module suppliers and trends in the PVmarket. Based on recent announcements of man~acturing expansions and technologyimprovements, the projected costs of PV modules are reasonable and maybe conservative.Future balance of system costs are more difficult to estimate, although as PV manufacturingaccelerates, it is reasonable to expect that the various BOS component costs will also decrease. Amajor uncertainty is the response of the international marketplace. As demand for PV systemsand components increases around the world, short-term prices may actually increase untiImandkturing capacity catches up to demand. In general, the projected costs for the dispatchablePV system are achievable.

Octoberf936 Page32

Debn8w8Pawef&L@tItcompeny PV:BOUUS Plmse f Find RqxuiCOOpemtiveAgmuwni No.DE-FCS6-SS~9

Task 3- Market Assessment

The major work items under this task are to assess the local, regional and national markets for thedispatchable PV peak shaving Systeu assess competing alternatives, and to assess market entryoptions. Market assessment work was done with the assistance of the Applied Energy Group(AEG). A draft of the final market research report was submitted on April 27. A final versiouincorporating Delmarva comments, was submitted in early June. In addition to the marketassessment work performed by AEG, the University of Delaware’s Center for Energy andEnvironmental Policy (CEEP) examined the economics of “multiple-use” systems in which thedispatchable PV system could be used for emergency and back-up power.

APPUED ENERGY GROUP MARKET RESEARCH

A surnnxuy of the results of the AEG market research is provided in this report. In general, all ofthe customer groups surveyed saw mnsiderable merit in both PV technology and the system asconfigured by Delrnarva. Based on AEGs researc~ there are virtually no concerns amongpotential Delmarva customers that PV technology will work mainly because of its inherentsimplicity. The modularity of the system was attractive because it can be usefil to a wide rangeof utility customers. Nearly all of the customers and trade groups interviewed agreed that peakshaving is a good application of PV technology because it offers an environmentally “fiendly”way to control utility bills. In addltio~ utility endorsement of any energy technology used oncustomer premises, including PV, is important. To utilities, the system can also be used fordistributed generation. However, for the vast majority of Dehnarva’s own commercial customers(and probably the customers of other utilities), a 2 to 5 year payback period is still the primarycriteria for making decisions about investments in energy systems and equipment. The cost of thetechnology therefore remains as the most significant barrier to its adoption.

A quantitative analysis helped to identi~ the size of the potential market. The marketquantification analysis did not attempt to take early adopters into account. Instead, the basicapproach was to examine the highest value of the dkpatchable PV system based on building loadshapes, electric rates and weather. This analysis focused on commercial business types (based onSIC codes) in three geographic tiers - Delmarva’s semice territory, the mid-Atlantic regio~ andthe entire country. The primary result of this analysis is an estimate of the number of commercialbuildings in each of the three tiers with the highest potential to adopt the system for economicreasons. Table 7 summarizes these results below.

As shown in Table 7, the size of the potential market for the dispatchable PV system is very largewithin the U.S. commercial sector. Awning a relatively low penetration rate of 5 percent of allof the high-potential adopters to account for technical limitations (inappropriate buildingorientatio~ unsuitable structure, shading, etc.) and other market acceptance factors, the market isstill substantial. At a 5 percent penetratio~ approximately 40,000 systems would be installed,requiring a minimum of about 400 MW of PV modules. In order to approach these fi~res, theinstalled cost must be reduced to between $3,100 and $7,700/kW. The upper end of this range isprobably achievable by about 1998. “

Ocrobef fm6 Pege33

Ddm8nmPuwor&ueJlcconIPmy—oPw@v04m-- No. DE-Fc36-93cnw69

W:B_ PhRSO1FhmlRep@

TABLE 7

SUMMARY OF MARKET RESEARCH RESULTS

Total Number of High PotentialRegion Buildings Adopters Percentage

Delmarva 27,520 1,980 14

Mid-Atlantic 1,157,405 215,003 19

Nation 5,876,079 818,317 14

Not withstanding efforts to reduce the cost of the systq the cost barrier can be overcomeinitially by increasing the value of the system ador by seeking out “early adopters.” Theseapproaches are described in greater detail below, in conjunction with additional information inAppendices A and B.

COMPETING TECHNOLOGIES AND VALUE-ADDED FUNCTIONS

As part of its scope of work under the PV:BONUS contract, the University of Delaware’s Centerfor Energy and Environmental Policy (CEEP) studied the value added to the dispatchable PVsystem when its use is expanded to include emergency lighting and backup power. Depending onthe application and the cost of competing alternatives, the value of a multi-fimctional system canbe as much as three times higher than a dispatchable PV system alone. The methodologyemployed and the detailed results are contained in AppendK B.

‘Ilk result has importantimplicationsfor additionalsystem development and marketentrystrategies. One of the highest valued service finctions is emergency power for computers. Basedon the costs of today’s UPS systems, the currentcost of dispatchablePV systems is nearlycompetitive. However, if a dual purpose PV system capable of being used for peak shaving andemergency service is to be made available, then considerable work is necessary to understand thismarket niche, along with the product requirements. While the system may be competitive withother emergency/backup systems, such as emergency lighting and UPS systems, it is not presentlycompetitive with conventional engine-generators. This means that wherever brief serviceinterruptions can be tolerated prior to starting an engine-generator, the engine generator willprobably be the pref~ed option.

5dma?wPowwa Ll@tcampany PWBONU3 Phisa +FhmlReportCOOperadveAgmunent No. DE-FC36-S3CHf#669

Task 5- Product Development and Testing Plan

The majorwork items underthis task are to develop manufacturingflows, assess technical andinstitutionalmilestones andto develop cost and other resource estimates for scaled upmanufacturing.

DEPLOYMENT AND TESTING PLANS FOR PHASE 2

Objectives of Phase 2 Work

Phase 2 of the PV:BONUS Program is an important step in refining and commercializing theDispatchable PV Peak Shaving System. There are several technical and business objectives whichestablish the foundation of the Phase 2 Statement of Work. These overall objectives aresummarized below:

1. Install and measurethe pefiormance of at least four DispatchablePV Peak Shaving unitsto validate both the conceptual and hardwaredesigns on a larger scale;

2. Ident@ technical improvements and fiwther cost reductions necessary for commercialintroduction in Phase 3;

3. Ident@ and analyze economic and policy strategies which fwilitate the commercialintroduction of the syst-,

4. Develop the strategic alhnces, manufacturing and support infrastructure necessary forproduction and installation; and,

5. Acquire the marketknowledge and develop the marketingplans necessary to focus on“earlyadopters.”

These objectives form the basis of specific tasks listed in the following summary of the Phase 2Statement of Work.

TASK la - Installation of Prototype Systems

Deknarva Power, with Ascension Technology and the AC Battery Corporation will Iiu-n.ishtheequipment and engineering services necessary for the instalMon of at least four (4) DispatchablePV Peak Shaving Systems. Initial contacts have already been made with prospective customers.Although final sites have not been selected, it is highly probable that at least three of the systemswill be installed at fwilities outside of the Delmama Power service territory.

Ocfdwr 19s6 Page36

De&mlw8Power4LigMcunlpany PV:BDNUSPbe 1find R@CoopwaffveAWUWIWMM. DE+C36-63CHW666

For each syste~ the following equipment and seMces will be provided to the purchasers:

Equipment

Services

10 kW PV ArrayAC Battery ModuleRoof Jack or Structural PV Array Support System ‘Monitoring EquipmentMiscellaneous Ekctrical Hardware

System Performance AnalysisSystem Design snd Installation PackageInstallation Supervision and Checkout

TASK lb - Development of a Grid-Independent Prototype System

In addition to the deployment of four dispatchable prototypes, a fifth grid-independent version ofthe system will be developed. This is in response to the market research analysis, which showedthat significant demand exists for packaged off-grid systems. While the largest share of thismarket is overseas, a significant domestic market also exists. The basic hardware developedunder this task can also be used for multi-fimction systems, capable of operating in either grid-connected or grid-independent modes. Systems such as these could be deployed as UPS oremergency power systems.

TASK 2- Refined Policy Analysis

Refined economic and policy analyses are needed to help efficiently focus fbture resources in thecommercialization effort. & a result of work in progress at the University of Delaware’s Centerfor Energy and Environmental Policy it is now possible to model the economic performance ofDispatchable PV Peak Shaving Systems under different ownership and policy scenarios. Inadditio~ the results of Phase 1 work indkated the importance of capturing “multiple use”benefits.

Under contract to Delrnawa Power, the Center for Energy and Environmental Policy at theUniversity of Delaware will perform a more comprehensive analysis of the economic value of aDispatchable PV Peak Shaving System with multi-iimction capabdities to utilities and customersincluding emergency Iight”hg and other back-up fimctions CEEP will also petiorm a cash flowanalysis from the perspectives of the utility and customer using a spreadsheet analysis tooldeveloped by CEEP.

TASK 3- Refined Market Research

The Phase 1 market research revealed significant potential “early adopter” market segments. Adetailed understanding of these early adopter segments is very important to the successfid

Octdw f996 Pege 36

De&nanmPowarCL@MC#nPIV W:BONU3 Phase f Find Report?vmnentNo. DE.FC3643ctuos69~Ag

commercializationof the DkpatchablePV Peak ShavingSystem. Penetration of the early adoptermarketswill help to build sales volume and reduce the productioncosts of the DispatchablePVSystem - mo critical steps necessary before entering the broader market. These segments includeFederal facilities (especially those managed by the General Services Administration), “nearoffshore” island electric utilties, and “green firms.”

Quantitative analysis done during Phase 1 using secondary data sources also indicated significantlong-term market potential. Direct market research data will help to refine the preliminary marketestimates, and to focus on the most important segments.

Additional market research performed under Phase 2 will include a telephone suxvey of 100potential “green firms” with follow-up visits to the top candidates to identfi financing options,pricing requirements, contracted requirements and other remaining barriers. In additio~ 500telephone-mail-telephone surveys will be administered in the mid-Atkmtic region to ident@ thenon-cost benefits to a broader range of potential customers of the Dispatchable PV Peak ShavingSystem

TASK 4- Solar Resource Building and System Performance Data Analysis

Because it is likely that the systems deployed in Phase 2 of PV:BONUS will be in areas outside ofDelrnarva Power% service territory, additional solar resource, building and system perfommncedata will need to be collected and analyzed. Using equipment installed under Task 1, data will becollected, analyzed and stored at a central location.

TASK 5- Battery~verter Sub-system Data Analysis and Development

Similar to Task 4, additional data must be collected and analyzed to evaluate the petiormance ofthe batteryliiverter system. Other enhancements to the batte@inverter module are alsoanticipated in Phase 2. Fdy, the resources necessary to scale up module assembly for theDispatchable PV Peak Shaving System must be identified. Under contract to Delmarva Power,the AC Battery Corporation will analyze batteryliiverter module data to provide a complete sub-system petiormance profile, including PV array size vs. PCS size, chopper performance, batterystorage capacity vs. discharge rate, and other performance parameters.

TASK 6- Development of System Integration, Marketing and Service Capability

The development of system integratio~ marketing and service capabilities are importantmilestones in commercializing the Dkpatchable PV Peak Shaving System. This task addressesseveral commercial and technical issues related to specifjirg sourcirg assembling and marketingthe system. The proposed sub-tasks will be performed in padlel with Task 1 in order to assurethat input from early customers is incorporated. These tasks include the development of array

Ocsobef199s Page 37

Ddmenm Poww c Lightcompany PV:BONUSPhase1Fhel Re#xntCooperativeAgreementNo.DE-FC36-93CHW69

specifications, investigation of PV module manufacturersandtechnologies and the localdevelopment of one or more assemblers. In additio~ a refined marketing plan will be developed.

The total estimated cost for the new work proposed in Phase 2 is approximately $1,300,000

Octo&r 1s96 Page38

APPENDICES

/ —

APPENDIX B

EMERGENCY POWER AS AN ADDITIONAL VALUE FROMDISPATCHABLE PV PEAK SHAVING

EMERGENCY POWER AS AN ADDITIONAL VALUE PROMDISPATCHABLE PV PEAK SHAVING

John Byrne, ContractorChandrasekhar Govindarajalu

Willett KemptonYoung-Doo Wang

Center for Energy and Environmental PolicyUniversity of Delaware

Prepared for PV-BONUS, Phase I Project:Ralph Nigro, Delmarva Power

Principal Investigator

July, 1994

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Customer needs foremergencypower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Emergency lighting 1Computer UPS 2Power for continued commercial operations 2

Comparison metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...2

Comparison to DPV-PS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...4

Relevance to Productand Marketing Decisions forDPV-PS . . . . . . . . . . . . . . . . . . 4

APPENDIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...7CASE I: Battery Powered Emergency Lighting Systems 7CASE II: Uninterruptable Power Supply for Computers (UPS) 8CASE III: Emergency Power for Continued Retail Operations 11

2

1. Introduction

The Dispatchable Photovoltaic Peak Shaving (DPV-PS) system configuration includesbattery storage to assure its dispatchability while maximizing its power at load peak. The batterystorage component of a DPV-PS system could additionally provide customers with on-site powerfor critical functions in the event of grid power failure. Most commercial buildings nowpurchase some form of emergency power, typically, emergency lighting. Since DPV-PS couldpotentially obviate or reduce the need for that purchase, it can increase the value of DPV-PSsystems to building customers.

Three emergency power options are reviewed below. All are widely used in thecommercial seetor. Each case examines typical system configurations and the value providedby current systems to customers. The functions of emergency power options are as follows:1. Emer~encv li~htinz. esweiallv li~htin~ in corridors and stairwavs. to E rmit exit from abuilding. Emergency lighting systems (ELS) are designed to switch on automatically duringpower outages, or during any local problem in the building that interrupts power.2. Uninternmtab]e Dower sum]ies (UPS) for computer owrations. Most of these are sized toprovide only enough energy storage to ride out short outages (approximately 10 to 30 minutes)or to permit orderly shutdown of network and file services.3. Power for continued commercial operations. Emergency power systems may be purchasedin order to allow normal operations, for example cash register sales or transaction processing,to continue throughout a power outage.

These three types of systems differ in the nature of customer needs being met, technicalrequirements, value, and economic cost. Metrics are developed for measuring the costs andvalue of each service and evaluating the possibilities of DPV-PS meeting these needs, includingrelative costs and potential opportunities or barriers as seen by customers.

2. Customer needs for emergency power

This seetion considers three emergency power applications in detail. The value of eachapplication is defined and the characteristics a DPV-PS must meet to satisfy customer needs aredelineated.

A. Emersencv lizhting Commercial buildings are required by code to provide lighting for safeoccupant exit in the case of power loss, primarily in halls and stairways. BOCA code requiresthat this lighting operate at least one hour, although available systems provide from 1.5 to 4hours. Typically, these are self-contained modular units, with charger, battery and lamps in onebox. Some systems have centralized batteries with conduit to the lighting. The centralizedsystems would be more adaptable to DPV-PS systems, which would have centralized batteries.Due to the need for cabling, centralized systems are most economical if designed into thebuildings initially. But DPV-PS systems could also replace modular units under certaincircumstances.

Three vendors were contacted (Brite-Lite, Lighting Design Center and United Electricsupply).

1

B. Commter UPS We consider both a PC-sized unit and the less common but higher-value,larger sized UPS systems for minicomputer or data center users.

Computer UPS can supply either a few minutes of power, intended to allow a rationalshutdown, or much longer, to allow continuing computer operations throughout a power outage.UPS systems are rated by load, number of minutes runtime, and switchover time (from gridfailure to emergency power activation). Protection from surges or spikes is another serviceadvertised and provided by many of the UPS units. Such protection may be required by onlya minority of customers, for example those in more rural areas.

Vendors contacted for this paper (American Power Conversion and Controlled PowerCompany) advertise very short switchover times (2-8 milliseconds). Advertisements from onevendor claim that “sensitive” electronic equipment can be shut down by an 8 msec blackout, butAC Battery’s research finds that almost all computers and electronic equipment can tolerate 5-8lost cycles (that is, 80-130 msec).

Some UPS systems come with software to shut down the computer system automatically.Some units are integrated, with a single box enclosing plugs and switches, battery charger andinverter, and the batteries themselves. Others have power conditioning and plugs in one box,and batteries in a separate box.

C. Power for continued commercial omrations. High-volume stores, or those providingemergency services, may invest in emergency power for business reasons. We examine in detailthe case of a retail grocery, whose investment in emergency power is justified by the high-volume, high-turnover nature of the business, and by the high costs incurred (in labor, loss ofperishables and customer satisfaction) if customers are forced to leave the store without ringingup their purchases. In the case examined, emergency power was provided by a natural-gasfueled generator. This provides emergency power at lower capital cost than batteries, butrequires up to several minutes to switch over to emergency power. A several-minute delay isacceptable in this case, as cashiers (and customers) wait briefly, then resume operations.

The Appendix provides more detail on the technology of each application, andquantitative analysis of the characteristics and costs.

3. ,Comparison metricsEach system we analyze is rated by kW and kWh. Whereas power generation equipment,

including PV, is typically compared on a $/kW basis, the cost of storage is rated by capacity.Thus, $/kWh is a good metric. The kWh ratings here are different than those of conventionalpower units, as they are based on the required energy storage of the unit per emergency use.

The required design energy storage depends upon the time during which the equipmentis intended to operate. For example, if a retail store has a 10 kW generator with a fuel tankcapable of running the generator 24 hours, but the store’s policy is to shut down after 3 hoursof power-off to protect against customers buying spoiled food, we calculate that energy systemat 10 kW x 3 hours (30 kWh), not its technical capacity of 10 kW x 24 hours (240 kWh). Thisrequired storage rating dictates the size of the battery component a DPV-PS system must include

2

to provide equivalent service. 1This approach allows us to compare different systems meeting divergent needs on

comparable metrics of kW, kWh, and capital cost per kW and per kWh. The following tablecompares the different types of systems and, within the emergency lighting type, comparesseveral different brands of emergency lighting. O&M costs for some units amounts to 30% ofthe O&M costs. Data and assumptions are spelled out in the more detailed cases discussed inthe Appendix.

TABLE 1Comparison of Energy Storage Costs for Emergency Power

System O&M Power Design kwh Switch- Capital CapitalType cost cost (kW) duration per over cost cost

($) ($tyr) (Hrs) Use time $lkW $/kWh

Emergency 35 5.8 0.005 3 0.016 0.1-10 6,480 2,160Lighting 4 sec(High)

Emergency 67 12.0 0.012 3 0.036 0.1-10 5,580 1,861Lighting o see(LOW)

Computer 799 103 0.315 0.33 0.105 2 msec 2,536 8,590UPS (Pc)

Computer 24,144 1,170 10 0.45 4.5 <8 2,414 5,365UPS (mini) msec

Generator 11,869 100 20.0 3 60.0 1-15 593 198min

Nlfita . C.. An-. A.. <*- Aatm*l”

The summary table indicates that current markets place a premium value on in-buildingemergency power, far above values common in utility calculations. There is also more than aten-to-one difference in prices charged within the emergency power market (admittedly, for verydifferent functions). The largest difference in cost is between a back-up combustion generator,at roughly $600 kW, and battery-based systems, at $2,500-$6,000 /kW. The cheaper generatortechnology is only possible at larger scales, and when a several-minute interruption ispermissible. The more expensive battery-based systems are required for small units (modular

1 Actually, DPV-PS battery storage may be sized above the required storage rating of an emergencypower application to assure the system’s ability to satisfy multiple objectives and/or to optimize storagefrom the PV component.

3

emergency lighting) and in applications requiring fast switchover time (computer UPS and, toa lesser degree, emergency lighting). Within the computer UPS and modular lighting categories,price differences are smaller, and primarily related to the size of the unit.

4. Comparison to DPV-PS

Can DPV-PS meet some or all of the functions of emergency power? There are bothequipment and marketing issues. Delmarva Power’s current PV-BONUS 1A unit can bemodified according to the manufacturer of the AC Battery module to serve the functionsdescribed here.2 If it were modified, there are still questions from the customer’s perspective,as to whether the DPV-PS package provides equivalent service at a competitive price.

One sizing issue is that the reservoir of usable power in a DPV-PS configurationfluctuates. For example, lead-acid battery electrochemistry limits depth of discharge in orderto prolong battery life. The amount of energy which can be drawn without any battery damageis called usable energy, and the total amount available if battery damage is acceptable is calledavailable energy. Normal operation of the DPV-PS unit anticipates discharging to a baselinedischarge. Even at this level, some usable energy remains for emergency power applications.3Since emergencies would not typically occur at the lowest state of charge, the typical emergencyenergy available would be much greater than the minimum guaranteed energy. Also, ifemergency power were needed at just the time of maximum discharge of DPV-PS batteries,added cost would accrue to the DPV-PS in the form of reduced battery life.

If maintenance of the DPV-PS system is provided as part of the total package, it offersan additional advantage. All dedicated emergency power systems require some form ofmaintenance. Battery-based systems require regular battery testing and battery replacementapproximately every 4 to 5 years. Bundling this testing and maintenance requirement into theDPV-PS system would be desirable to the building customer.4

5. Relevance to Product and Marketing Decisions for DPV-PS

All three emergency power functions reviewed here could potentially provide added

2 Robert Flemming, AC Battery, personal communication, June 1994.

3 This terminology and information was provided by AC Battery,

4 Economic analyses conducted to date on the peak-shaving application designed by the Center forEnergy and Environmental Policy and Delmarva Power has assumed utility maintenance. See, e.g.,Byrne et al., 1993, “Commercial Building Demand-Side Managment Tools: Requirements forDispatchable Photovoltaic Systems, “ in l’he ConferenceRecord of the Twenty l’birdIEEE PhotovoltaicSpecialists Conference, Louisville, Kentucky and Byrne et al., 1994, “PV-DSM as a Green InvestmentStrategy, Proceedings of the Fifih National Conference on Integrated Resource Planning, NationalAssociation of Regulatory Utility Commissioners, Washington, DC.

4

opportunities for marketing DPV-PS. Some of the marketing advantages are identified below.

A. Emerpencv Li~hting. Modular ELS units are sold at the same high costs as computer UPSsystems, but are very easy to install and maintain. Furthermore, their distributed nature makesthem robust even to major building disruptions (e.g., fire in the service room). On the otherhand, modular installations require battery testing and replacement for a dispersed set of lightingunits. This can add considerably to the O&M cost. DPV-PS can be modular in operation bydistributing at least a portion of the battery bank to different sites in a building and connectingthe dispersed storage to separately wired PV panels.

However, a centralized emergency lighting system would seem to be a more logicalmatch for DPV-PS. Such a system would have centralized battery storage with conduit toindividual lights. However, this would be most economical in buildings for which the DPV-PSsystem is being offered during the design phase, or a replacement of an existing centralized ELS.Retrofitting a centralized DPV-PS-powered lighting system to a building already fitted withmodular units could be difficult.

Marketing an ELS function as an added benefit of DPV-PS seems the most straightforward of the three emergency power applications investigated here. ELS does not involvecomplex design or service criteria, as computer UPS does (see Appendix), and has an establishedmarket because it is mandated by all states. Service guarantee could be similar to that providedby current manufacturers.

Market potential can be rated as promising.

B. Commter UPS. This is a high-value application which can be met by the DPV-PS system(with minor equipment changes).

The dual demands of power from DPV-PS—for peak-shaving and emergencypower—will require further thought and testing with customers. However, potentially it couldbe sold under a guarantee of supplying sufficient power (e.g., 1-5 minutes) to permit orderlyshutdown. In many cases, it should be possible for computer UPS service from DPV-PS toallow continuous operations for extended periods during a power outage. It maybe worthwhileto conduct an analysis of the timing of commercial power outages compared with peak shavingneeds, to arrive at a figure for the probabilityy that only minimum storage would be available atthe time of an outage. More complex controls could be built in; for example, overriding thepeak-shaving mode of operation when a UPS order is given by a customer. Similarly, thecustomer might have an override switch allowing the DPV-PS system to use all available batterypower, at the cost of reduced battery life or even battery damage, in order to meet extendedemergency computer UPS needs. Such options would increase the value of the DPV-PS systemby providing additional emergency power when the customer is willing to pay for it, an optionnot provided by separate UPS systems.

Another issue may be vendor credibility. Network administrators or MIS managers(Management Information Systems) may perceive a company which specializes in UPS to bemore credible than a vendor of a PV power unit. It would help if the vendor wereconspicuously identified as a division of the electric utility, since electric utilities generally havehigh credibility in such technical matters. This is certainly a solvable problem, since some

utilities have already entered the business of selling or leasing UPS systems to their customers.sMarket potential is probably less strong than ELS but well worth additional investigation.

C. Power for Continued Commercial Oerations. The case we examined was a retail groceryusing an internal-combustion generator. As long as the customer can tolerate the longerswitchover times, the generator option is inexpensive and offers long duration for continuedoperations (energy is stored more economically in fuel than batteries). Therefore, the DPV-PSwould offer little added value if it replaced a generator.

Nevertheless, the market for continued commercial operations is diverse. Some potentialcustomers would not want a generator due to noise, pollution, hazard to employees from rotatingmachinery, etc. As with computer UPS systems, for this function some customers might findit attractive that they were guaranteed a certain minimum period of power supply, but that alonger time would be available in most cases.

More generally, customer functions which would benefit from having uninterruptiblepower should be further investigated as part of a DPV-PS marketing strategy. These mayinclude services not now provided, and that the customer may not have considered to this point.Uninterruptible power can be a clear benefit that increases the value of DPV-PS systems andwhich can draw support from additional managers in a customer’s organization.

Further study is needed before market potential of this category of emergency power canbe fairly assessed.

S In Florida, Tampa Electric has leased over 100 UPS units to its customers. Baltimore Gas andElectric is currently providing UPS systems to its customers.

6

APPENDIX

CASE I: Battery Powered Emergency Lighting Systems

We describe the characteristics of two types of emergency lighting systems, centralizedand modular. Lighting costs can be accurately estimated from manufacturer data for modukirsystems but not for centralized ones, so we provide cost data only for modular units.

Case IA: Centralized Emer~encv Li~htinp SvstemsCentralized emergency lighting systems have a centralized battery and charger. The

lighting load is powered by this unit in case of power outages. The emergency lightingrequirement itself is determined based on the building layout. Costs can be established only withreference to a specified building design. The centralized battery and inverter capacity of theemergency lighting system could be replaced by the battery and power conditioning unit whichform a part of the DPV-PS system.

Case IB: Modular Emerzencv Lizhtin~ SvstemsEmergency lighting systems are also available as independent units comprised of lamps,

battery and charger unit. Typically, these modular units are rated at 6V and 12V and have two6 W lamps as light sources. 12 V units are typically used in industrial establishments. In caseof power failure, this system is designed to provide light for about 3-4 hours. Data collectedfrom various retailers is tabulated below. The capital cost comparisons per kWh are based onthe costs paid for typical usage time capacity of 3 hrs. We estimate operating and maintenancecosts as a $25 expenditure once every five years for battery replacement. The batteries usuallycarry a 3 year warranty.

TABLE A-2Cost Comparisons of Modular EmerEencv LiEhtinE Svstems. --- -.

Manufacturer System Bulb Capital Capital Design hourscost($) Wattage( Cost in cost in of use

w) $/kWh $/kW

Brite-Lite 6V 49.0 6W 2,722.0 8,160.0 3 hrs

Brite-Lite 12V 75.0 25 W 1,077.0 3,000.0 3 hrs

Lighting 67.0 12 w 1,861.0 5,580.0 3 hrsDesign Center

United Electric 35.0 5.4 w 2,160.0 6,480.0 4 hrs.supply

7

CASE II: Uninterruptable Power Supply for Computers (UPS)

We consider two size ranges of computer UPS systems: PC or network server, andminicomputer. These span a 300 W to 20 kW range (one that is likely to be met by a DPV-PSsystem of the type now contemplated).

Case 11A: Personal Comouters or PC Network ServersA PC-size UPS can be purchased when any single office worker performs computer

functions that are considered critical. Considering an office building as a whole, a typical sizemight be 50 PCs on network and 3 PC-type file and print servers. In the PC-size market,American Power Conversion is the largest manufacturer. The example building could be fittedwith three UPS units—three American Power Conversion “Smart-UPS 900” systems (no UPSon individual workstations) and a criterion of twenty minutes of emergency power beforeshutdown.

We assume that the DPV-PS battery and power conditioning completely replaces the UPSsystem, rather than using the battery of the DPV-PS in conjunction with the UPS power-conditioning and switching. This means that the DPV-PS system would have to be capable offast switchover, and the DPV-PS vendor would have to convince the customer that the systemwould reliably provide “uninterruptibility”.

Our contact manufacturer provided no guarantee on uninterruptible power. It was statedthat there was no way to distinguish customer errors (e.g. overloading) from failures of UPSequipment. They did provide a $25,000 guarantee if customer equipment was damaged bysurges or spikes.

Unit specification: APC Smart-UPS 90@’● List price $799 (discounts available)● Three units typically installed for servers on a 50 PC network● Each unit powers a single IBM PS/2 80 file server● Units draw 3/10 amp, 36 watts, standby (without charging) (at $0.09/kWh, running costs

are $30/year)(larger battery unit increases only to 0.45 amp)

● Maximum AC output 630W @ 0.7 power factor (900 VA) can run 7 minutes● Half power, 315W @ 0.7 power factor can run 20 minutes (note that the relationship

between runtime and power output is nonlinear)(thus, 0.074 kWh of energy at max, 0.105 kWh at half power)

● Batteries in 900 unit: 4 x 6V, 10 amp-hour batteries; $77 replacement cost for set of 4(thus, 0.24 kWh total battery storage per unit—more than twice the output energy)

Alternative unit specification: APC SmartCell External Battery Supplement● To increase storage, one can add external battery packs (chain up to 10)

6 Data from manufacturer (dominant in PC UPS market): American Power Conversion, WestKingston, RI (800) 800-4272. Information also fkom technician at 1-800-788-2409.

8

● Each single external battery pack is 4 cells, 12V, 17 amp-hour, total cost $599 list● Thus, cost per kWh DC is $737 for smart cell, more than twice the cost of 900 battery.

Operating and maintenance costs, single 900 unit:● Manufacturer states 3.5-4 years to replace battery, we take 4● Battery cost at $77, plus 1 hour time of system manager ($20), total $97 or $24/year● Standby electricity is 36 watts; if always plugged in, 315 kWh/year @ $0.07/kWh =

$22/year● Assume one unit failure per 7 years, requiring $400 repair = $57/year● Thus, total O&M cost is $103/year (battery replacement represents 20% of O&M)

Summary for a single UPS 900:Key assumptions: running at 1/2 load; 20 minutes availableat315WPower: 0.315 kWEmergency Power Time: 20 minutesDelivered energy: 0.105 kWhO&M cost: $103/year (estimated from battery +repairs+electricity)System cost: $7609/kWh; $2536/kWBattery cost: $733/kWh; $2441kW

Case IIB: UPS for Larger ComwterLarger computer UPS systems would be sized for machines such as a VAX or small IBM

mainframe, or for a data center including a central computer and peripheral equipment. Themanufacturer contacted for this case (Controlled Power Company) has a complete line of largecomputer UPS systems. Their single-phase units range from 2.17 kW to 20 kW. We reviewboth ends of this range here. They also have a line of three-phase systems, 120/208 volt output,ranging up to 500 kVA.7

For this manufacturer, we estimate O&M costs based on the cost of a five-year servicecontract; this is the equivalent of out-sourcing O&M. Of the variety of service contractsoffered, the ones quoted here are comprehensive; they are priced on a five-year basis andinclude battery replacement when needed. This manufacturer provides no guarantee tocompensate for data loss, only repair or replacement of their equipment within the warrantyperiod.

The. next larger commercial systems would use different technology: pulse-widthmodulated, double-conversion (input is always converted to DC and back to AC, even duringnormal power conditions). These systems are available in the 10 kVA to 500 kVA range, single

7Informationfrom ControlledPower Company, 1955StephensonHwy, Troy, MI 48083,1-800-642-9624.

9

and three-phase. 8Regarding technical specifications, they advertise switchover time as less than one-half

cycle (8 msec). They advertise “high” AC to AC efficiency of “up to 95%”. Since the powerto the computer always runs through the UPS, we assume that this represents a 5% loss duringcomputer operations. To calculate the ongoing energy costs, we assume that the computer loadis 1/2 the maximum UPS load rating, electricity is 8 Q/kWh with no demand charge, and thatminicomputers or mainframes would be operational 24-hours a day, 365 days per year. If theseassumptions are correct, for the two larger systems, the energy costs of UPS are significant--40% as much as the O&M costs.

The table compares three models, spanning the range of sizes of single-phase units.Calculations of cost per energy unit are based on half-load usage, since running these units atfull load dramatically shortens time available (for example, the HV 14000 runs 16 minutes at halfload, but only 6 minutes at full load, so the half-load condition provides 33% more energy).

TABLE A-1Cost Comparisons of Single-Phase Computer UPS Systems

Model Full load Minutes Capital O&M Energy Capital CapitalkW @ 1/2 cost $/year cost cost cost

load $/year $/kW $/kWh

MD31OO 2.17 32 min @ $3,695. $190. $38 3,421. 6,415.1.08kW

HV14000 10 16 min @ $11,470. $430. $175 2,294. 8,602.5 kW

HV25000 20 27 min @ $24,144. $820. $350 2,414. 5,365.10 kW

8 Manufacturers for the larger commercial UPS systems are Liebert (Columbus, Ohio) and ExideElectronics (Raleigh, North Carolina). Industrial applications use even larger uninterruptable powersupplies, for example, for control rooms or maintaining processes. A manufacturer of these systems isWeiss Instruments in Philadelphia (1-610-647-4650).

10

CASE HI: Emer~encv Power for Continued Retail Operations

The commercial establishment used as an example for this application is a grocery store(Acme) of 1,500 square meters, having 11 cash registers. The main electrified end-uses in thisstore are refrigeration, air conditioning and lighting.

The store is assumed to have a natural gas-based emergency backup system, whichautomatically switches on in case of power failure. It is assumed to take several minutes to startup. This system can supply 20 kW (three-phase) or 13 kW (single-phase) load at 60 Hz and 0.8power factor. It is sized to meet the energy requirements of cash registers and lighting systems.There is no energy backup for refrigeration and air conditioning systems.

The manager of the store was not aware of the capital cost or operating costs of thegenerator. During the past year, the generator was used only two times, once for one hour andonce for two hours. When asked about use of PV systems for meeting emergency requirementsof the store, he felt it would depend on the system’s cost-effectiveness.

The generator set consists of a multi-fuel capability engine (natural gas or gasoline), abrushless AC alternator, a control panel, voltage regulating system, a cooling system and a base.The automatic transfer switch which could be used along with the generator set has a utility-to-generator option that enables retransfer of load in 0-30 minutes.9 This time delay is notsufficient to maintain power to computers achieved by UPS systems.

Specifications of the internal-combustion generator setMake: OnanW Capacity: 20 WDesign hours of use: 3(estimated)Delivered Energy: 60 kWh (based on 3 hours)Cost of generator plus switch :$11,869.00Operation and maintenance costs: estimated $100/year maintenance and fuel consumption of 9.0m3/hr when running, no fuel or electric use when not operatingCapital Cost: $ 198/kWh, $ 593/kW

9Contact for data on technical and economic characteristics of the gas generator: Eastern Generators,1-800-397-1983.

11

APPENDIX C

ASSESSMENT OF POLICY, REGULATORY AND INSTITUTIONAL BARRIERS TO THECOMMERCIALIZATION OF A DISPATCHABLE PV PEAK SHAVING SYSTEM

ASSESSMENT OF POLICY, REGULATORY AND INSTITUTIONAL BARRIERS TOTHE COMMERCIALIZATION OF

A DISPATCHABLE PV PEAK-SHAVING SYSTEM

John Byrne, Sub-ContractorYoung-Doo WangWillett KemptonThulasi Kumar

Center for Energy and Environmental PolicyUniversity of Delaware

Prepared for PV-BONUS, Phase I ProjectRalph Nigro, Delmarva Power

Principal Investigator

July 1994

Table of Contents

I. Typical Barriers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...’.... . . . . . 1

II. Options for DPV-PS Commercialization .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

11-1. Rate Adjustment Option .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .H-2. Shared Savings Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .II-3. Lease-to-Own Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .II-4. Green Partnership .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

IV. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v. Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v-1. Energy Policy Act of 1992 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .V-2. Regulatory Reforms . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

55“66

8

9

11

1114

2

This paperPV peak-shaving

identifies typical barriers hindering the commercialization of a dispatchablesystem (hereafter DPV-PS)

barriers. A utility-customer partnership iscommercialization under existing constraints.

1. Typical 13am”ers’

iiIIU ULXXXILES U~LIUIl> lUI

argued to be a major

.—s1 --_u L ..-— .: ---c-- overcoming thosetool for speeding

Major institutional issues bearing on the commercialization of utility DPV-PS systemsinclude financing issues, owner-tenant differences, technical risks and uncertainties, buildingcode problems, aesthetic concerns, the lack of a DPV-PS system infrastructure, and the lack ofinformation by many of the parties involved. These issues are briefly reviewed below.

● DPV-PS involves high capital costs which increase the purchase price of a newbuilding in which it is installed. Builders, and especially those building forspeculative sale, seek to minimize construction costs and may, therefore, judgeDPV-PS uneconomical DPV-PS as a retrofit building technology may, likewise,face financial hurdles. Buldings tenants are unlikely to be willing to shoulder thelong-term debt represented by an investment in DPV-PS, and building ownersreap few benefits unless demand (kW) elements of their electricity operating costsare significant and borne by them (rather than passed onto, for example, tenantsin the building.

● Interest rates charged by existing lending institutions on loans for buildingeuipment are typically higher than those charged for new construction or buildingpurchase. The unavailability of low-interest loans is a disincentive to purchaseand install a PV-DSM system with a higher front end cost, even though it couldlower operating costs in the long run.

● A major barrier to DPV-PS commercialization is the demand, often, for highrates of return by commercial customers (typically, 5 year paybacks or less aresought for efficiency investments-- see Arthur D. Little, 1993; Greely, Harris andHatcher, 1990). Additionally, energy costs are in many cases a small fraction ofoperating costs for businesses. Only high rates of return justify, in suchinstances, attention to what will be comparatively modest gains to the firm.

‘Detailed discussion of barriers to renewable energy and DSM options can be found in Nadel(1993), Moskovitz (1992 and 1993), Arthur D. Little (1993) and Hirst and Brown (1990).

1

● Most commercial utility customers are tenants. But investment in energyefficiency is from two to five times more likely when the utility customer is thebuilding owner (Hines, 1990). Furthermore, high turnover rates among tenantsare a problem. A survey of 143 commercial sites found 45 changes of occupancyover a two year period, representing 16% per year (Peterson, 1990). Sometimesa building owner refuses to let a tenant install energy-saving equipment eventhough the tenant would pay 100% of the cost due to a concern that maintenanceworkers would not be able to work with the new equipment.

● Investing in new PV-DSM systems maybe perceived as risky. Users of DPV-PSsystems cannot be sure whether these systems will perform as projected withoutmajor maintenance problems. Such concerns have been found to be veryimportant to decision makers in all end-use sectors. Risk aversion can affectcustomer participation in utility DPV-PS programs and might be more importantto some than the technology’s costs and benefits.

● Building codes and standards are mostly concerned with health, safety andreliability. End-use efficiency is seldom addressed. The efficiency advantage ofDPV-PS must therefore compete with mandated building equipment andimprovements for customer investment. When building owners must carry outmandated actions, the available capital for discretionary investments (such asefficiency upgrades) is reduced and, consequently, opportunities for DPV-PS arediminished.

● DPV-PS investments are affected by code and standard requirements which relateto height and wind loads, resulting in higher design and mounting structure costs(Arthur D. Little, 1993).

● The spatial requirements for PV energy collection can raise aesthetic questionsthat impact DPV-PS commercialization. Physical limits on collector area andlocation for some buildings may also narrow the market potential for thistechnology. Additionally, basic questions surrounding solar access remainunsolved. This places DPV-PS at an institutional disadvantage in the buildinginvestment market.

● The scale of commercialization of DPV-PS technology will almost certainly belimited in the near term by the modest number of firms and professionals skilledin the design, system integration, engineering, financing, operations andmaintenance, and construction of DPV-PS systems. In addition, m,mpetence andexpertise in the manufacture, distribution, and servicing of DPV-PS systems isweak at this time. In general, DPV-PS infrastructure problems are likely tohamper rapid, large-scale commercialization of the technoogy at least for the next2-5 YUUS.

2

● Credible information on the performance of DPV-PS technologies is very limited.Such information is critical to those who decide on the commercial deploymentand adoption of this new technology, including investors, regulators, buildingowners/tenants and utilities. In particular, information regarding the technicaland economic viability of PV-DSM technologies is often scarce, and the absenceof such data can be a barrier to the adoption of such systems.

11. Options for DPV-PS Commercitdizuh”on

By enhancing the overall economic competitiveness of renewable energy, thecommercialization of DPV-PS can be promoted. To bolster private investment in DPV-PS,federal, state or local governments could reduce private risk or increase private returns throughtax incentives, energy-efficiency building codes, regulatory reforms, and investment in R&D andprocurement. Likewise, reduction in subsidies to conventional energy sources2 and therecognition of social and environmental costs in conventional energy prices would establish amore accurate market in which DPV-PS would have to prove its competitiveness.

The Public Utility Regulatory Policies Act (PURPA) of 1978 and the Energy Policy Act(EPACT) of 1992 constitute the basic policy framework for renewable energy development inthe U.S. Key elements of this framework include: the requirement that utilities evaluatedemand- and supply-side options at their avoided costs; the institutionalization of integratedresource planning; the provision of tax credits for renewable energy investment; and regulatoryconsideration of social and environmental costs of electricity consumption. While these andother policy milestones have been put in place, these are unlikely to mobilize an open andcompetitive market for DPV-PS systems (see Appendix for regulatory and policy reforms relatedto DPV-PS applications). The barriers identified in the previous section will continue to thwartefforts at early development of DPV-PS. But new technologies have often had to compete onan uneven playing field. The central question at this juncture is whether and under whatconditions PV applications have the best chance of overcoming these barriers.

Not only domestic (U. S.) but also international energy markets have important roles toplay in DPV-PS commercialization. European and Japanese manufacturers and markets. aremajor influences on PV development. Multilateral lending organizations, including the WorldBank and regional development banks, are and will play a significant role in shaping energyefficiency and renewable energy use in developing countries. These international efforts couldopen up PV markets, reduce PV system costs and, consequently, enhance PV commercialization.

20ne estimate places the federal subsidy to U.S. energy industries at more than $44 billion as of1984 (FIavin and Lenssen, 1992).

3

In general, assessments of the near- to mid-term prospects of PV commercialization arepositive. A typical, optimistic view is that of the Worldwatch Institute (1990) which notes thatPV cost was 339C/kWh in 1980, 30$/kWh in 1988, and estimated to be 10C/kWh in 2000 and4C/kWh in 2030. A less optimistic, but nevertheless positive assessment was recently renderedby Arthur D. Little (1993) which specifically identified utility applications as among the mostpromising.

PV’S future is clearly filled with opportunity: but the current cost of 20-25 Q/kWh -underscores the challenge to effective commercialization. PV technology needs to gathermomentum through innovative market penetration strategies. The commercialization of DPV-PScan be accelerated either through a decline in unit costs and/or a targeting of high-valueapplications. By penetrating high-value niche markets, PV-entreprenuers will be better able tocompete while also contributing to the expansion of PV production facilities. In turn, expandingproduction capacity drives down costs to open up broader PV markets:

Given current PV costs and technical constraints and under the present regulatory andinstitutional constraints, the most plausible marketing strategy for dispatchable peak-shaving PVis to form a utility-customer partnership and target the non-residential building sector as the“early” market. The arrangement typical of partnerships appears to fit well with the conditionsthat are necessary for the promotion of PV in the utility sector. The reasons are as follows: 1.)neither customers nor utilities can justify investing in DPV-PS systems because costs exceedbenefits (Hoff and Wenger, 1992); 2.) if the combined benefits of dispatchable peak-shavingcapacity to utilities and moderation of electricity costs to customers during the peak-demandperiod are weighed against system costs, DPV-PS is much closer to cost-effectiveness thanexpected (Byrne et al., 1994); and 3.) commercial building sector applications offer high valueto utilities and customers because the electricity market of this sector is highly sensitive to peakdemand costs.

This section focuses on partnership arrangements which foster a closer relationshipbetween customers and utility than typical DSM programs. In the case described below,building owners who are also utility customers are targetted to avoid the frequent turnover ratesand owner-tenant conflicts discussed above. The two parties enter into a partnership to sharethe costs and capture the benefits of DPV-PS applications. Four basic forms of partnership areexamined: 1.) a partnership stimulated by a rate adjustments package offered by a utility; 2.) ashared savings partnership in which the benefits of peak shaving are shared between utilities andtheir customers; 3.) lease-to-own strategy which allows partners to share risks until one party

%le current world market for PV is around 50 MW, growing 20 to 30% annually in which theUnited States supplies one-quarter of total production (Kozloff and Dower, 1993).

4The Utility Photovoltaic Group (UPVG) has adopted the concept of forward pricing in itsTEAM-UP Proposal to the U.S. Department of Energy. The strategy follows the logic of developinghigh-value markets in order to stimulate increased production and, hopefully, lower unit costs.

4

is prepared to own and operate the system themselves; and 4.) a Green Partnership whichutilities and customers share costs and benefits via a utility-managed limited partnership.

11-1. Rate Adjustment Option

Electric utilities across the country seek to induce customers to participate in specificDSM programs by offering them incentives often in the form either of rebates or lower electricrates. Customers must undertake specific actions to improve end-use efficiency such as theinstallation of higher SEER-rated HVAC equipment, in order to qualify for utility incentives.The electric utility may provide a rebate on the purchase of such equipment or reclassify thecustomer making them eligible for a lower electric rate. Rate adjustments can include: lowerenergy charges, lower demand charges, or demand charges based on the peak demand reducedby the DSM system. In almost all cases the customer is responsible for providing most of thenecessary investment and has ownership and responsibility for the maintenance of the equipment.

The adjustment of rates or the provision of rebates frequently is targetted to commercialand industrial customers. An increasingly common focus of these programs is to stimulate fasterpurchase of load shifting or peak load shaving equipment. In return for the long-term benefitsat reduced risk provided the customer through rate adjustments or rebate, the utility expects thecustomer to contribute a portion of the capital cost of the system.5 DPV-PS could beencouraged by this mechanism. But its high initial cost may narrow the number of commercialcustomers who would invest in the technology on this basis.

II-2. Shared Savings Option

Often, consumers might be able to improve the energy efficiency of their operationsthrough investment but are either unable or unwilling to undertake the necessary capitalexpenditure. This can be due to the relatively high cost of capital to the customer (comparedto, for example, a utility), higher perceived risk, lack of information, or other factors. In suchcases, shared savings partnerships with an energy service company (ESCO) can be attractive.Typically, these partnerships involve the energy service provider financing all, or most, of thecapital investment for energy efficiency improvements. This investment is then recovered, witha profit, by the investor receiving a portion of the energy bill savings of the customer. In manycases, these ESCOS are non-regulated subsidiaries of electric utilities.b

The primary determining factor for many shared savings contracts appears to be the sizeof the initial investment. Small investments typically require shorter periods for their recovery(3-5 years), requiring a smaller percentage of the bill savings received by the investor (30-50%).

5The maximum customer’s contribution would be equal to the net present value of the stream ofbenefits (at lower risk) accrued during the useful life of the equipment.

6A notable development in this area is that the regulatory environment is evolving to allow util itiesto enter into such partnerships (Eto et al., 1992).

5

Large investments typically have recovery periods from 15 to 20 years and 50 to 80 percent ofbill savings may be needed by the investor to justify such a high outlay of capital.

The shared savings can overcome the high first-cost problem of PV-DSM, especially ifthe utility is the investor. This is because the utility has direct access to bill savings and alreadyhas a well-established service relationship with the customer. The likely arrangement would befor a utility to propose to purchase the majority of the necessary investment and offering to sharea portion of the resulting bill savings on the basis of the marginal benefits it derives from usingPV-DSM instead of a generation option. The payback period would probably be the expectedlife of the equipment, and the share ratio (proportion of savings withheld) would be between 70and 80 percent at current PV prices. This would require a long-term relationship with thecustomer which a utility would be capable of. But customers may prefer arrangements whichdo not presume long pay backs.

II-3. Lease-to-Own Op[ion

To address the long payback issue, shared savings contracts can specify that the customeror utility has the option to buy out the interest of the other after a certain period. Contractscould stipulate that the customer or utility receives ownership of the equipment by paying aspecified amount from a schedule that is based upon the number of years that the contract hasbeen in effect. This would create the basis for a lease-to-own partnership. In this partnership,the customer eventually can become the sole owner after a period of time, allowing the initialinvestor (the utility) to recover its capital contribution by being paid a lump sum. In itsapplication to DPV-PS, the lump sum transfer from customer to utility reflects the remainingvalue of the utility’s initial investment,

11-4. Green Partnership

The Green Partemership7 is an innovative investment strategy in which customers andutilities contribute to a green investment fund for the purpose of purchasing a DPV-PS systemmuch in the same manner that “green pricing” joins both parties in a common effort to makeadvgmce purchases of technologies that are currently not cost-effective to the individual parties(Moskovitz, 1992 and 1993). Both strategies are intended to encourage early sales of newtechnology in the hope that such sales will stimulate more rapid price reductions by renewabletechnology manufacturers.

In green pricing, the customer agrees to pay a premium rate in order to enable moreutility investment in renewable energy sources. - Me-transfer to an investment fund most of the potential bill

Green Partnership asks customers tosavings from the DPV-PS system and

‘For fiwther information on this strategy see Byrne et al. (1994). “PV-DSM As a GreenInvestment Strategy. ” Proceedings of Fifih Narional Conference on Integrated Resource Planning.Kalispell, Montana. May 15-18, 1994.

6

most of the tax savings from the investment. The remaining cash savingsit may be attractive to customers who are interested in promoting the etechnology as part of their corporate image. The highly-visible nature offor some customers, be more important than the magnitude of the econort

Under the Green Partnership, the utility would contribute an amounland distribution benefits that it would derive from the dispatch of a peakTypically, these benefits would take the form of avoided costs; i.e., postponand/distribution upgrades. For the utility to contribute its avoided costspossible to sell the capacity and energy freed up by the operation of the D]market exists for the freed capacity and energy, the utility might decline to [costs to the partnership in order to avoid potential lost revenues (namely, thave earned by selling electricity before the DPV-PS system was put int{likelihood would limit utility participation to those companies experiencing

In this arrangement, the customer transfers to the Green Partnershitax and bill savings that could be realized from the operation of the Paddition, the utility contributes its avoided costs for the credited capacity

There is, however, a necessary condition of regulatory refornimplementation of the Green Partnership. Typically, state regulators utilCost (TRC) test to evaluate utility-sponsored DSM programs. 9 Undelbetween utilities and their customers are generally not counted as benefits

In the case of the Green Partnership, this would mean that thecontributed by the customer would not figure into the cost-benefit analysithat such savings should be recognized as partnership transactions outside thservice regulatory framework. In effect, the Green Partnership COU”unregulated affiliate transaction, in which case only the avoided costcontribution to the Partnership by the utility would be subject to regulatorytreatment were extended to the proposed Green Partnership, technology vto achieve market penetration in the near term.

8This does not represent double-counting if the avoided cost contributiondispatch function of the PV-DSM system, while the potential bill savings to the cfor the periods when the system is ~t under a utility dispatch requirement (i.e.,Partnership incurs the debt associated with the purchase of the DPV-PS system.

%ere are several approaches used by electric utilities and regulatorseffectiveness of utility-sponsored DSM programs. These approaches differ main]and benefits considered in the calculation of net present values or benefit-cost Icommonly defined in terms of their impacts on participants, non-participants, 1(Krause and Eto, 1988).

7

III. Conclusion

DPV-PS systems facea wide array of policy, regulatory and institutional barriers thatdisadvantage the technology in the electricity market place. But many of these barriers aresimilar in form to those that affected adversely the eeonomics of earIier innovative technologies.Which reforms in the regulatory and policy areas are appropriate and useful to assuring a fairmarket evaluation of DPV-PS systems, there are speeific options that can be implemented byutilities to spur commercialization in the absence of reforms.

These options involve investment partnerships between utilities and seleeted customersthat utilize already established contractual relations (e.g., the Rate Adjustment and SharedSavings Partnership discussed above). But in addition, new partnerships options (especially, theGreen Partnership) offer utilities and selected customers the opportunity to be early developersof this technology by spreading risks and pooling benefits. When reforms subsequently occur,these early adopters will be ideally positioned to take advantage of this promising PV market.

8

Iv. References

Byrne, John, Steven Letendre, Ralph Nigroand Young-Doo Wang. 1994. ’’PV-DSMASa Green Investment Strategy. ” Proceedings of F@hNationalConference onlntegratedResource Planning. Kalispell, Montana. May 15-18, 1994.

Center for Environmental Legal Studies (Pace University). 1990. Environmental Costsof Electricity (New York: Oceana Publications).

Eto, J., A. Destribats and D. Schultz. 1992. Sharing the Savings to Promote EnergyE@ciency. Lawrence Berkeley Laboratory, Energy and Environment Division.

Flavin, Christopher and Nicholas Lenssen. 1992. “Policies for a Solar Economy. ”Energy Policy 20(3).

Greely, Kathleen M., Jeffrey P. Harris and Ann M. Hatcher. 1990. “Measured EnergySavings and Cost-Effectiveness of Conservation Retrofits in Commercial Buildings. ”Proceedings of ACEEE 1990 Summer Study on Energy E@ciency in Buildings 3:95-108.

Hines, V. 1990. “Study Shows Conservation Trends in D.C. New Construction. ” EnergyUser News 15(10): 4.

Hirst, Eric and Marilyn Brown. 1990. Closing the E@ciency Gap: Barriers to theEflcient Use of Energy, Resources, Conservation and Recycling. Oak Ridge NationalLaboratory.

Hoff, Tom and Howard J. Wenger. 1992. Photovohaics as a Demand-Side ManagementOptions: Concept Development (San Ramon: Pacific Gas and Electric Company).

Kilmarx, Mary and Peter Wallis. 1991. “DSM LCP.” DSM Quarterly. Fall: 22-25.

Kozloff, K. Lee and Roger C. Dower. 1993. A New Power Base: Renewable EnergyPolicies for the Nineties and Beyond (Washington, DC: World Resources Institute).

Krause, Florentin and Joseph Eto. 1988. Least-Cost Utility Planning Handbook for PublicUtility Commissioners (Washington, DC: National Association of Regulatory UtilityCommissioners).

Little, Arthur D. 1993. Z?ieMarket for Solar Photovoltaic (PV) Technology. EPRI TR-102290. March.

Moslcovitz, David. 1992. Renewable Energy: Barriers and Opponunities; Walls andBridges. A paper for the World Resources Institute. The Regulatory Assistance Project.

9

Moskovitz, David. 1993. Green Pricing: Experience and Lessons Learned. TheRegulatory Assistance Project.

Nadel, Steven M. 1993. “UtilityDemand-SideM anagement Experience and Potential:ACritical Review.’’ LBLAdvance IRPSeminar. Vol.1.

Nadel, Steven M., Michael W. Reid and David R. Wolcott (eds.). 1992. Z?egulato~Incentives for Demand-Side Management (Washington, DC: ACEEE).

Peterson, Frances J. 1990. “Remodel and Tenancy Changes: Threats to the Reliabilityof Commercial Conservation Savings. ” Proceedings of ACEEE 1990 Summer Study onEnergy Eficiency in Buildings 3:165-172.

Rowe, John. 1990. “Making Conservation Pay: The NEES Experience. ” 7%eElectricityJournal 3.

U.S. Congress. 1992. Energy Policy Act.

10

v. Appendix

As of 1991, U.S. electric utilities were spending nearly $2 billion per year, representingapproximately 1% of their revenues, on demand-side management (DSM) investments. Someutilities were investing 7% of their gross revenues in DSM (Kilmarx and Wallis, 1991). Sucha DSM commitment is stimulated to a large degree by policy and regulatory changes. More than20 states have adopted or experimented with regulatory reforms since 1989 to promote DSMoppotunities.

The passage of the Public Utility Regulatory Policies Act (PURPA) of 1978 was a majorcontributor to changes in the regulatory environment. Integrated Resource Planning (IRP), alsocalled Least-Cost Planning (LCP), grew out of PURPA and is now well-established in manyutilities as an alternative to traditional supply-side planning. The nature of conventionalregulation had created a bias in favor of traditional supply-side resources. IRP creates a levelplaying field in the acquisition of DSM and supply-side resources. The federal reform of utilityregulation with PURPA stimulated a variety of state-level reforms.

State-level DSM reforms date back to at least 1980 when Washington State legislationdirected that utilities be granted a 2% bonus rate of return on the equity portion of DSMinvestments and allowed utilities to submit conservation program expenditures as investments,similar to supply-side investments. ERAM, the Electric Revenue Adjustment Mechanism, wasproposed in 1981 in California as a means of decoupling the link between base revenue and thelevel of sales. In 1986, the Wisconsin Public Service Commission adopted a performance-basedincentive scheme. 10

Togehter, these federal and state policy activities established an institutional basis for aDSM market. Without these actions, it is difficult to imagine DSM as having achieved anythinglike its current market status. Additional federal and state initiatives are identified below whichunderscore the importance of policy/regulatory reform and the focus and direction of theelectricity market.

v-1. Energy Policy Act of 1992

The 1992 Energy Policy Act (EPACT) has several provisions promoting utility DSMprograms (including DPV-PS when it effectively acts as an efficiency upgrade to a building’soperations-- see Hoff and Wenger). EPACT promotes IRP and also includes such provisionsaddressing DSM incentives, energy-efficient mortgages, performance contracts and thecommercialization of energy-efficient products to enhance energy end-use efficiency of theelectricity sector. Key provisions of EPACT are summarized below:

‘%e mechanism was designed in such a way that one percentage point additional return on theequity portion of DSM would be received for each 125 MW of demand reduction.

11

●

●

States must conduct hearings on the merits of requiring electric utilities to employIRP, a process that would include comparing the life-cycle costs of conventionalpower plants, energy conservation and efficiency, power purchases and renewableenergy sources. [$11 l(a),(d)]

States must consider establishing rate structures for regulated electric utilities toencourage investments in “appropriately monitored” energy efficiency,conservation and demand-side management. [~11l(a)]

DOE is authorized to grant up to $250,000 to state regulatory authorities toencourage electricity DSM measures, including conservation, efficiency and loadmanagement, and DSM for meeting natural gas supply needs. [~112]

National Energy Policy Plans submitted by the President in 1993 and afterwardshall include a least-cost energy strategy prepared by DOE. The strategy is to setgoals and priorities that promote energy efficiency, renewable energy and otherenergy technologies that reduce greenhouse gas emissions. [$1602]

Competitive DOE grants totaling $10 million for FY94 and $50 million for FY95are authorized for helping federal agencies meet the Act’s energy efficiencyrequirements. Agencies are permitted and encouraged to participate in utilityincentive programs for gas, electricity, and water conservation, and to negotiateincentives with utilities. [~153]

DOE is authorized to participate in the development of revolving funds for energyefficiency projects in state and local government buildings (up to $1 million perstate). To qualify, a state must have “demonstrated a commitment to improvingenergy efficiency, ” by, for example, adopting energy codes (specified in SubtitleA) or by raising at least 75 percent of the fund’s capital from non-federal sources.[~141(a)]

If appropriations are available, DOE is directed to make payments of 1.5 G/kWh,adjusted for inflation after 1993, for electricity produced by new qualifiedrenewable energy facilities. Facilities eligible for the payment are state-ownedgeneration plants or electrical cooperatives, and must begin operating within thenext 10 fiscal years, and each facility may receive payments for only 10 fiscalyears after starting operation. [$1212]

Beginning on January 1, 1993, incentives given to residential customers byregulated public utilities for installing energy conservation measures is excludedfrom gross income for tax purposes. “Energy conservation measures” are definedby reference to the National Energy Conservation Policy Act of 1978 (P.L. 95-619). For utility incentives to commercial and industrial customers, the exclusion

12

begins on January 1, 1995, and is limited to 40 percent of the incentive for 1995,50 percent for 1996, and 65 percent after 1996. [$1912]

● The 10 percent business tax credit for solar and geothermal equipment is extendedindefinitely. The credit for solar and geothermal equipment was introduced in1978 as part of President Carter’s National Energy Plan. [$1916]

● The Departments of Housing and Urban Development (HUD) and Agriculture(USDA) are to issue energy efficiency standards for residential buildings financedthrough federal mortgage programs. New buildings must satisfy these standardsto receive federal financing assistance. [$lOl(c)]

● EPACT promotes homogeneous building energy efficiency codes by requiringstates to begin upgrading their residential and commercial building energyefficiency codes [$101(d)].”

● A task force on energy efficiency and private mortgages, established in 1990 bythe Cranston-Gonzalez National Affordable Housing Act, is to makerecommendations on notifying potential home buyers about energy-efficientmortgages (EEMs), which are designed to provide incentives for purchases ofresidences with major energy efficiency measures installed. HUD mustimplement and promote a pilot five-state EEM program for existing residences.The pilot program must be expanded to new construction and the remaining stateswithin two years, unless HUD reports to Congress why expansion would beimpractical. [~105, $106]

● Federal agencies are permitted to enter into multiyear energy-savings performancecontracts, subject to certain specified requirements. Under such contracts, privatefirms may pay to install energy-saving equipment in federal buildings in returnfor a share of the future energy cost savings. DOE is to establish procedures foruse by agencies based on specified criteria. [$155]

● DOE is to establish a program to install energy conservation measures in federalfacilities and federally assisted housing. Agencies may submit proposals forfunding under this program. DOE is to study the potential of using federalpurchasing power to promote the development and commercialization of energyefficient products, and study potential cost-effective energy savings achievable infederal buildings. [$152(h)]

llFor residential buildings, states must certify that their code meets or exceeds the 1992 Councilof American Building Oftlcials (CABO) Model Energy Code (MEC). For commercial buildings, statesmust certify that their code meets or exceeds ASHRAE/IES Standard 90.1-1989 (Standard 90.1).

13

V-2. Regulatoq Reforms

Several regulatory reforms made thus far for utility DSM programs can contribute toDPV-PS systems as well. These include: 12

● Decoupling utility profit from sales through such mechanisms as the ElectricRevenue Adjustment Mechanism (ERAM)13 has been an important regulatoryreform for removing existing disincentives against utility DSM investments,thereby encouraging the successful implementation of IRP.

● Someutilities are allowed to recover from customers the revenue losses associatedwith approved DSM measures. In other cases, like New England Power, therates are determined annually and demand-side activities are accounted forexplicitly in future-test-year forecasts, all but eliminating the problem of lostrevenues.

● Currently, prudent supply-side resource costs can be fully recovered in utilityrates. To promote utility investments in DSM, some utilities are allowed torecover DSM costs on terms equivalent to those for the recovery of supply-sidecosts. In other cases, participating customers pay the full cost of utility-sponsored DSM programs through a surcharge.

● Removing disincentives to DSM is a necessary, but not sufficient condition tomake DSM attractive to at least some utilities. For some utilities, performance-based incentives have been offered with impressive results (Rowe 1990).14

● Shared-savings incentives have been approved in 13 states*5 and the District ofColumbia. Shared-savings incentives are usually accompanied by guarantees ofprogram cost recovery and by a decoupling mechanism that remove disincentivesassociated with reduced sales.

**SeeNadel et al eds 1992. Regulatory Incentivesfor Deman&Side Management for a more.detailed discussion of sev;ral of these items.

*3ERAM adjusts base rates to ensure that an electric utility collects its full authorized revenuerequirement, despite its level of sales. ERAM was main]y intended to remove a perceived anti-DSMbias.

*41nthe case of New England Electric System (NEES), DSM incentives have been successful,giving the incentives movement new credibility. In 1991, NEES allocated a greater portion of its budgetto DSM than any other investor-owned utilities in the country, about 5% of revenue.

IsThese states include California, Georgia, Idaho, Indiana, Maine, New Hampshire, New Jersey,New York, Ohio, Oregon, Rhode Island, Vermont and Wisconsin.

14